Hepatitis C virus infection of cholangiocarcinoma cell lines

Hepatitis C virus (HCV) infects the liver and hepatocytes are the major cell type supporting viral replication. Hepatocytes and cholangiocytes derive from a common hepatic progenitor cell that proliferates during inflammatory conditions, raising the possibility that cholangiocytes may support HCV replication and contribute to the hepatic reservoir. We screened cholangiocytes along with a panel of cholangiocarcinoma-derived cell lines for their ability to support HCV entry and replication. While primary cholangiocytes were refractory to infection and lacked expression of several entry factors, two cholangiocarcinoma lines, CC-LP-1 and Sk-ChA-1, supported efficient HCV entry; furthermore, Sk-ChA-1 cells supported full virus replication. In vivo cholangiocarcinomas expressed all of the essential HCV entry factors; however, cholangiocytes adjacent to the tumour and in normal tissue showed a similar pattern of receptor expression to ex vivo isolated cholangiocytes, lacking SR-BI expression, explaining their inability to support infection. This study provides the first report that HCV can infect cholangiocarcinoma cells and suggests that these heterogeneous tumours may provide a reservoir for HCV replication in vivo.


INTRODUCTION
Hepatitis C virus (HCV) is an enveloped positive strand RNA virus classified in the genus Hepacivirus of the family Flaviviridae. Worldwide, approximately 170 million individuals are persistently infected with HCV that leads to a progressive liver disease, including cirrhosis and hepatocellular carcinoma (reviewed by Scheel & Rice, 2013). The major cell type in the liver supporting HCV replication is hepatocytes (Kandathil et al., 2013;Wieland et al., 2014). Hepatocytes and cholangiocytes derive from a common epithelial progenitor cell that proliferates during liver inflammation (Roskams, 2006); however, to date there are no published studies investigating the permissivity of cholangiocytes to support HCV infection.
Cholangiocarcinomas account for approximately 10 % of all primary hepatic cancers and can be classified as intrahepatic or extrahepatic in location (Patel, 2006;Roskams, 2006). Cholangiocarcinomas are heterogeneous and can arise from cholangiocytes, liver progenitor cells (Komuta et al., 2012) Chen, 1998), highlighting a potential new reservoir that merits further investigation.

RESULTS
Cholangiocarcinomas support HCV pseudotype particle (HCVpp) entry To investigate the ability of cholangiocytes to support HCV entry we screened primary cells isolated from 10 donor liver explants with various disease aetiologies, together with cell lines derived from normal tissue (H69), intra-(CC-LP-1 and CC-SW-1) and extra-hepatic cholangiocarcinomas (Sk-ChA-1 and Mz-ChA-1). We confirmed that all of the cholangiocarcinoma derived lines expressed the epithelial markers EpCAM, CK19 and epithelial membrane antigen NCAM and GCTM-5. HCVpp failed to infect any of the primary cholangiocytes, however, we observed efficient entry into two cell lines, CC-LP-1 and Sk-ChA-1 (Fig. 1a). All cell lines tested supported vesicular stomatitis virus G pseudotype particle (VSV-Gpp) entry, demonstrating functional lentiviral promoter activity in these cells (Fig. 1a). To investigate whether cholangiocarcinoma cells supported the entry of HCVpp expressing diverse envelope glycoproteins, we generated pseudoparticles expressing E1E2 glycoproteins cloned from HCV genotype 1a/b acutely infected subjects (Osburn et al., 2014). As controls we included the well-characterized Huh-7 hepatoma line and the non-permissive claudin-1 null human embryonic kidney 293T cell line. All HCVpp strains infected CC-LP-1 and Sk-ChA-1 cholangiocarcinoma lines with comparable efficiency to Huh-7 but failed to infect 293T (Fig. 1b). We confirmed that CC-SW-1 and Mz-ChA-1 cells were refractory to all patient-derived HCVpp (data not shown). To investigate the receptor dependency of HCVpp infection of the permissive cholangiocarcinoma lines, we assessed the ability of anti-CD81 and anti-SR-BI antibodies to inhibit HCVpp (strain H77) infection. Both antibodies inhibited HCVpp entry, demonstrating receptor-dependent entry (Fig. 1c). In addition, anti-E2 (3/11) and polyclonal IgG purified from chronic-HCV infected subjects inhibited HCVpp infectivity ( Fig. 1c) but had no effect on VSV-Gpp infection (data not shown). These studies show that some tumour-derived cholangiocytes but not those isolated from non-tumour liver tissue support HCVpp entry.

Cholangiocarcinoma express HCV entry factors
To investigate HCV entry factor expression in vivo we stained cholangiocarcinoma liver tissue from two donors with antibodies specific for CD81, SR-BI, claudin-1, occludin and epithelial marker CK19. Cholangiocarcinoma from both donors expressed all four HCV entry factors, albeit with low CD81 expression (Fig. 2a), whereas biliary epithelia from the normal non-tumour margin lacked SR-BI expression (Fig. 2b). To assess whether the cholangiocarcinoma cell lines show a similar profile of receptor expression to the tumour tissue, the cells were stained for receptor expression along with Huh-7 hepatoma cells as a positive control. The permissive cell line Sk-ChA-1 expressed all four entry factors at comparable levels to Huh-7 hepatoma cells (Fig. 3a). Of note, CC-LP-1 cells expressed CD81, SR-BI and occludin; however, we failed to detect any claudin-1 expression (Fig. 3a). Both permissive cell lines expressed CD81 and occludin at the plasma membrane; however, claudin-1 was predominantly intracellular in Sk-ChA-1 cells and not observed in CC-LP-1 cells (Fig. 3b). The two nonpermissive cholangiocarcinoma lines, CC-SW-1 and Mz-ChA-1, expressed low levels of SR-BI, similar to that observed for biliary epithelia in non-tumour liver tissue, suggesting that this may be the limiting factor for HCV entry. These data show that cholangiocarcinoma and epithelial cells isolated from the tumour express all four HCV entry receptors, consistent with their permissivity to support HCV entry.
To determine whether HCVpp expressing patient-derived envelope glycoproteins required claudin-1 to initiate infection we assessed their ability to infect the claudinnull cell line, 293T, and CC-LP-1 cells following claudin-1 overexpression. Claudin-1 expression was confirmed by flow cytometry (data not shown). HCVpp strains only infected 293T cells expressing claudin-1, demonstrating claudin-1 dependent entry. However, the same viruses infected parental CC-LP-1 cells, however, their infection levels were increased following claudin-1 expression in this cellular background (Fig. 4). In summary, diverse HCVpp strains infect CC-LP-1 cholangiocarcinoma cells independent of claudin-1, -6 or -9 expression.
Cholangiocarcinoma Sk-ChA-1 support HCV entry and genome replication Sk-ChA-1 cells support cell-culture-derived HCV (HCVcc) (strains SA13/JFH-1 and JFH-1) replication as assessed by enumerating NS5A expressing cells, whereas CC-LP-1 failed to express detectable levels of viral antigen or RNA. Similar results were obtained with CC-LP-1 cells transduced to express claudin-1, suggesting that their non-permissivity to support HCVcc replication was not due to a claudin-1 independent viral uptake pathway. Foci of NS5A expressing Sk-ChA-1 only comprised on average 2-4 cells, suggesting minimal viral spread. Infection was inhibited by polyclonal patient IgG from pooled HCV-infected donors, neutralizing anti-CD81, Telaprevir (VX-950) and interferon-a (Fig. 5a). To compare the permissivity of Sk-ChA-1 cells to support HCV replication to primary human hepatocytes (PHHs) we measured HCV RNA levels 72 h post-infection and included permissive Huh-7 cells as a positive control. We noted comparable levels of HCV RNA in Sk-ChA-1 and PHHs from two donors (Fig. 5b). Huh-7 cells supported significantly higher levels of HCV replication, most likely  (Osburn et al., 2014) for CC-LP-1 and Sk-ChA-1 along with Huh-7 and 293T control cells. (c) Anti-receptor antibodies (anti-CD81 or SR-BI), anti-HCV Ig or anti-E2 (3/11) inhibition of HCVpp infection of CC-LP-1, Sk-ChA-1 and control Huh-7 cells. Data are presented relative to control antibody. N53 independent experiments. ****P,0.0001, ***P,0.001, **P,0.01, *P,0.05. explained by their negligible Toll-like receptor 3 (TLR3) expression and limited ability to sense replicating RNA . Attempts to infect the highly permissive Huh-7.5 cell line with extracellular virus secreted from Sk-ChA-1 or PHHs cells failed to establish infection, most likely due to the low frequency of infected cells and comparable to the life cycle observed in infected neuroepithelioma cells (Fletcher et al., 2010).
Several cellular factors have been demonstrated to facilitate HCV infection and replication, including miR-122 (Jopling et al., 2005), Niemann-Pick C1-like 1 (NPC1L1) cholesterol absorption receptor (Sainz et al., 2012) and Apolipoprotein E (ApoE) (Jiang & Luo, 2009). We therefore investigated the expression of these factors in the two cholangiocarcinoma cell lines. Neither CC-LP-1 nor Sk-ChA-1 cells expressed detectable levels of miR-122 (Fig. 5c). Both cell lines expressed NPC1L1 at similar levels to Huh-7 cells, as measured by Western blotting (data not shown), suggesting that neither of these reported host-cell factors explained the differential permissivity of the cholangiocarcinoma cells to support HCVcc infection. In contrast, we noted a significant difference in ApoE expression between the two lines, with the permissive Sk-ChA-1 secreting detectable levels of ApoE (Fig. 5d). These results demonstrate that Sk-ChA-1, but not CC-LP-1, cholangiocarcinoma cells support HCV replication.

DISCUSSION
CC-LP-1 and Sk-ChA-1 cells derived from intra-and extrahepatic cholangiocarcinoma tissue, respectively, support HCV entry or replication. In contrast, HCV did not infect primary cholangiocytes. Infection of cholangiocarcoma cells was inhibited by antibodies specific for CD81, SR-BI, E2 glycoprotein and by pooled immunoglobulin from chronic HCV infected patients. Of note, CC-LP-1 expressed negligible levels of tight junction protein, claudin-1 mRNA or protein and yet supported the entry of HCVpp expressing a range of patient-derived envelope glycoproteins. CCl-LP-1 cells do not express detectable levels of claudin-6 or -9 mRNA suggesting that infection is claudin-independent. We confirmed that the infectivity of all HCVpp strains tested were claudin-1 dependent in 293T cells and expression of claudin-1 in CC-LP-1 cells significantly enhanced HCVpp entry. To the best of our knowledge this is the first report of claudin-1 independent HCV entry into hepatic derived epithelial cells and supports a model where CD81 and SR-BI mediate viral attachment and internalization in the absence of claudin-1, -6 or -9 co-expression.
Cholangiocytes and hepatocytes arise from a common progenitor cell type, termed oval cells in rodents and reactive ductular cells in humans (Roskams, 2006). During chronic liver diseases, reactive ductular cells become activated and differentiate into cholangiocytes and hepatocytes, depending on the nature of the liver injury. Cholangiocarcinomas develop from cholangiocytes and reactive ductular cells, however, recent studies provide evidence that intrahepatic cholangiocarcinoma can also originate from hepatocytes (Fan et al., 2012;Sekiya & Suzuki, 2012). This occurs through activation of NOTCH and AKT signalling in hepatocytes, leading to the conversion of hepatocytes into cholangiocytes with concurrent malignant transformation (Fan et al., 2012). These data may explain why viral hepatitis is a risk factor for cholangiocarcinoma, since infected hepatocytes can undergo NOTCH activation and lineage conversion (Iwai et al., 2011;Sekiya & Suzuki, 2012).
We did not assess the ability of purified hepatic progenitor cells to support HCV infection, however, NCAM and EpCAM-positive cholangiocytes isolated from donor livers with primary biliary cirrhosis are known to contain an enriched population of reactive ductular cells, and were refractory to infection. In addition, primary cholangiocytes isolated from subjects with primary sclerosing cholangitis, a condition that predisposes to cholangiocarcinoma, were refractory to infection (Roskams et al., 1990) (E. Gershwin, personal communication). CC-LP-1 and CC-SW-1 cells are derived from intrahepatic cholangiocarcinomas (Shimizu et al., 2006) whereas Sk-ChA-1 and Mz-ChA-1 are derived from extrahepatic tumours of the biliary tree and gall bladder, respectively (Knuth et al., 1985), indicating that the ability to support HCV replication is not restricted to intrahepatic cholangiocarcinomas. Viral tropism is defined at multiple levels of the virus life cycle, including entry, RNA replication and assembly (reviewed by Scheel & Rice, 2013). In addition to the four essential factors for HCV entry, CD81, SR-BI, claudin-1 and occludin, several additional factors facilitate infection. These include epidermal growth factor receptor (EGFR) (Lupberger et al., 2011), Niemann-Pick C1-like 1 cholesterol absorption receptor (Sainz et al., 2012) and the liverspecific micro-RNA, miR-122 (Jopling et al., 2005). Virus particles are secreted in association with apolipoproteins and ApoE expression enhances viral infectivity and virus particle production (Da Costa et al., 2012;Hueging et al., 2013;Jiang & Luo, 2009). Sk-ChA-1, but not CC-LP-1, cells supported HCV replication that was inhibited by interferon-a, anti-CD81, HCV-positive pooled patient IgG and Telaprevir. While the levels of infection observed in Sk-ChA-1 cells was significantly lower than Huh-7 cells, Sk-ChA-1 cells supported similar levels of infection to PHHs. Huh-7 cells have previously been shown to support significantly higher HCV replication than PHH, likely due to reduced sensing of HCV RNA and a lack of TLR3 expression in these cells (Farquhar & McKeating, 2008;Wang et al., 2009). Since Sk-ChA-1 but not CC-LP-1 cells support HCV replication, we quantified the expression of additional host factors reported to facilitate infection to explore the differences in these cell lines to replicate HCVcc. None of the cell lines expressed miR-122, similar to our previous observations with blood-brain barrier endothelial cells and HepG2-CD81 hepatoma cells that support low level HCVcc replication (Fletcher et al., 2012;Israelow et al., 2014). Of note, only Sk-ChA-1 cells secrete detectable levels of ApoE, which, together with their expression of the essential HCV entry factors, may explain their permissivity to support HCV replication.
In summary, we have identified two cell lines derived from cholangiocarcinoma tissue that support efficient HCV entry and low-level HCV replication. Our data raises the possibility that cholangiocarcinomas may represent a reservoir for HCV infection in vivo and warrant further studies to establish the role of HCV in cholangiocarcinoma pathogenesis. (a) Sk-ChA-1 cells were inoculated with HCVcc strain SA13/JFH (titre of 10 6 IU ml "1 based on Huh-7 cells) in the presence or absence of anti-HCV Ig (100 mg ml "1 ), anti-CD81 antibody (clone 2s131, 10 mg ml "1 ), protease inhibitor Telaprevir (1 mg ml "1 ) and interferon-a (10 IU ml "1 ). Data are presented as f.f.u. ml "1 calculated as NS5A antigen-expressing cells. (b) Sk-ChA-1, but not CC-LP-1, cells support HCV SA13/JFH RNA replication at comparable levels to PHH. HCV RNA levels were normalized to Huh-7 cells. (c) Cholangiocarcinoma cells do not express miR-122 but control miR-210 was detected. (d) Sk-ChA-1 cells secrete low levels of ApoE, whereas we failed to detect ApoE from CC-LP-1 cells, where the dotted line indicates the ELISA cut-off point. N53 independent experiments.
Liver tissue and confocal imaging. Formalin fixed and paraffin embedded biopsies were obtained from patients with cholangiocarcinoma that was diagnosed according to standard biochemical and histological criteria: all tissues studied were selected by an experienced histopathologist. Liver sections (10 mm) were deparaffinized and rehydrated in water followed by low temperature antigen retrieval. Sections were blocked with 2 % Caesin (Vector Laboratories) and incubated with anti-CD81, anti-SR-BI, anti-claudin-1 and antioccludin along with anti-CK19. Bound antibodies were detected with Alexa-conjugated secondary anti-species antibodies and labelled sections mounted using Fluorescent Imaging Media (Dako). Images were acquired using an upright Zeiss 780 laser scanning confocal microscope (10061.4NA objective), where microscope settings were optimized for each fluorescent protein to obtain the highest signal to noise ratio whilst controlling for cross talk. Background fluorescence intensities were determined from the fluorescent signal of an Ig isotype control.
Flow cytometry. Cell surface receptor expression was monitored by live-cell staining and flow cytometry as previously reported (Fletcher et al., 2012). Briefly, cells were incubated with anti-SR-BI, anti-CD81 or anti-claudin-1 for 1 h. To detect occludin cells were fixed and permeabilized followed by incubation with a primary antibody. After a brief wash the cells were labelled with a fluorescent conjugated secondary antibody for 1 h. Thereafter, cells were fixed with 1 % paraformaldehye (Sigma) and data collected using a FACS calibur flow cytometer (BD Biosciences) and analysed with FlowJo software (Tree Star).
HCVpp and HCVcc genesis and infection. Pseudoparticles were generated by transfecting 293T cells with plasmids encoding a human immunodeficiency virus (HIV) provirus expressing luciferase and vesicular stomatitis virus G (VSV-G), a panel of HCV envelope glycoproteins (Dowd et al., 2009) or a no-envelope control, as previously reported (Hsu et al., 2003). Supernatants were harvested at 48 h post-transfection, clarified and filtered through a 0.45 mm membrane. Virus-containing medium was added to target cells plated in 96-well plates seeded at 5610 5 cells cm 22 . At 72 h post-infection, cells were lysed and luciferase activity measured in a luminometer (Lumat LB 9507). HCVpp infectivity was calculated by expressing the HCV or VSV-G luciferase signal (relative light units, RLU) relative to the No env RLU value. HCVcc NS5A-positive foci were enumerated and infectivity expressed as f.f.u. ml 21 .
To generate HCVcc, plasmids encoding chimeric SA13/JFH (Jensen et al., 2008) or J6/JFH-1 (Lindenbach et al., 2005) were used to generate HCV RNA as previously described (Lindenbach et al., 2005). Briefly, RNA was electroporated into Huh-7.5 cells, supernatants collected at 72 and 96 h and stored at 280 uC. Various cell lines and PHHs were inoculated with HCVcc for 6 h in the presence or absence of anti-HCV Ig (100 mg ml 21 ), anti-CD81 (2s131 at 10 mg ml 21 ), Telapravir (1 mg ml 21 ) or interferon-a (10 IU ml 21 ). Unbound virus was removed by washing and the cells re-fed fresh media plus or minus antiviral agents and propagated for 72 h before fixing for NS5A detection or extraction of total cellular RNA for HCV RNA quantification. Cells were fixed with ice-cold methanol and stained for NS5A with mAb 9E10 and an isotype-matched Alexa 488-conjugated anti-mouse IgG2a.
Neutralization of HCV infection. Huh-7, CC-LP-1, Sk-ChA-1 or 293T cells were seeded in 96-well plates at 5610 5 cells cm 22 . The cells were incubated 24 h post-seeding with 10 mg ml 21 anti-receptor or irrelevant IgG control mAb. After 1 h, HCV-H77pp, VSV-Gpp or No-envpp, or HCVcc, was added and incubated for 72 h at 37 uC. In addition, anti-E2 mAbs or HCV + IgG was incubated with virus for 1 h prior to infecting the appropriate target cells. At 72 h postinfection, luciferase activity was measured for HCVpp infections, or cells stained for NS5A. The percentage neutralization was calculated relative to the irrelevant IgG control.
Transduction of cells to express claudin-1. CC-LP-1 or 293T cells were transduced to express claudin-1 as previously described (Flint et al., 2006;Harris et al., 2010). Briefly, packaged lentiviruses to express claudin-1 were generated by cotransfection of 293T cells with plasmids encoding VSV-G protein, HIV Gag-Pol, and pTRIPclaudin-1 (1 : 3:3 ratio). Cells were seeded at 4610 5 cells cm 22 and infected 24 h later with the packaged lentivirus. After 12 h, cells were seeded into appropriate plates either for HCVpp or HCVcc infection as described.
Real-time reverse transcriptase PCR. RNA was prepared using the Qiagen RNeasy or MiRNeasy kit for microRNA analysis. Purified cellular RNA samples were amplified for HCV RNA (Primer Design Ltd), claudin-1, -6 or -9, or miR-122 in a quantitative reversetranscription PCR (qRT-PCR) in accordance with the manufacturer's guidelines (CellsDirect kit; Invitrogen) using an ABI7500 PCR machine (Applied Biosystems). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) or miR-210 were included as endogenous controls for amplification efficiency, and HCV amplification normalized to GAPDH using the DDCt method.
ApoE ELISA. Cells were cultured for 24 h and supernatant harvested. Secreted ApoE levels were measured using a commercial ApoE ELISA (Abcam) according to the manufacturer's instructions.
Statistical analysis. Results are expressed as the mean±1 standard deviation of the mean. Statistical analyses were performed using Student's t-test in Prism 6.0 (GraphPad) with a P,0.05 being considered statistically significant.