Long-term propagation of serum hepatitis C virus (HCV) with production of enveloped HCV particles in human HepaRG hepatocytes


  • Ndiémé Ndongo-Thiam,

    1. Centre de Recherche en Cancérologie de Lyon (CRCL), INSERM U1052/CNRS UMR5286, Lyon, France
    2. Université Claude Bernard Lyon 1, Lyon, France
    Search for more papers by this author
  • Pascale Berthillon,

    1. Centre de Recherche en Cancérologie de Lyon (CRCL), INSERM U1052/CNRS UMR5286, Lyon, France
    2. Université Claude Bernard Lyon 1, Lyon, France
    Search for more papers by this author
  • Elisabeth Errazuriz,

    1. Centre Commun d'Imagerie Laennec (CeCIL), Lyon, France
    Search for more papers by this author
  • Isabelle Bordes,

    1. Centre de Recherche en Cancérologie de Lyon (CRCL), INSERM U1052/CNRS UMR5286, Lyon, France
    2. Université Claude Bernard Lyon 1, Lyon, France
    Search for more papers by this author
  • Sylvie De Sequeira,

    1. Centre de Recherche en Cancérologie de Lyon (CRCL), INSERM U1052/CNRS UMR5286, Lyon, France
    2. Université Claude Bernard Lyon 1, Lyon, France
    Search for more papers by this author
  • Christian Trépo,

    1. Centre de Recherche en Cancérologie de Lyon (CRCL), INSERM U1052/CNRS UMR5286, Lyon, France
    2. Université Claude Bernard Lyon 1, Lyon, France
    3. Hospices Civils de Lyon, Hôpital de la Croix Rousse, Service d'Hépatologie et de Gastroentérologie, Lyon, France
    Search for more papers by this author
  • Marie-Anne Petit

    Corresponding author
    1. Centre de Recherche en Cancérologie de Lyon (CRCL), INSERM U1052/CNRS UMR5286, Lyon, France
    2. Université Claude Bernard Lyon 1, Lyon, France
    • CRCL INSERM U1052/CNRS UMR5286, 151 Cours Albert Thomas, 69424 Lyon Cedex 03, France
    Search for more papers by this author
    • fax: +33 472 681 971

  • Potential conflict of interest: Nothing to report.

  • Supported by INSERM and by a grant from Agence Nationale de Recherche sur le SIDA et les Hépatites Virales (ANRS). N.N. was supported by a PhD fellowship from the ANRS


HepaRG human liver progenitor cells exhibit morphology and functionality of adult hepatocytes. We investigated the susceptibility of HepaRG hepatocytes to in vitro infection with serum-derived hepatitis C virus (HCV) particles (HCVsp) and the potential neutralizing activity of the E1E2-specific monoclonal antibody (mAb) D32.10. The infection was performed using HCVsp when the cells actively divided at day 3 postplating. HCV RNA, E1E2, and core antigens were quantified in HCV particles recovered from culture supernatants of differentiated cells for up to 66 days. The density distributions of particles were analyzed on iodixanol or sucrose gradients. Electron microscopy (EM) and immune-EM studies were performed for ultrastructural analysis of cells and localization of HCV E1E2 proteins in thin sections. HCV infection of HepaRG cells was documented by increasing production of E1E2-core-RNA(+) HCV particles from day 21 to day 63. Infectious particles sedimented between 1.06 and 1.12 g/mL in iodixanol gradients. E1E2 and core antigens were expressed in 50% of HCV-infected cells at day 31. The D32.10 mAb strongly inhibited HCV RNA production in HepaRG culture supernatants. Infected HepaRG cells frozen at day 56 were reseeded at low density. After only 1-3 subcultures and induction of a cell differentiation process the HepaRG cells produced high titer HCV RNA and thus showed to be sustainably infected. Apolipoprotein B-associated empty E1E2 and complete HCV particles were secreted. Characteristic virus-induced intracellular membrane changes and E1E2 protein-association to vesicles were observed. Conclusion: HepaRG progenitor cells permit HCVsp infection. Differentiated HepaRG cells support long-term production of infectious lipoprotein-associated enveloped HCV particles. The E1E2-specific D32.10 mAb neutralizes the infection and this cellular model could be used as a surrogate infection system for the screening of entry inhibitors. (HEPATOLOGY 2011;)

Hepatitis C virus (HCV) infection is a major health problem worldwide because at least 70% of infections persist and cause chronic hepatitis, which may progress to liver cirrhosis and hepatocellular carcinoma.1 The lack of robust cell culture and small animal models remain stumbling blocks to HCV research. Fortunately, thanks to the identification of a full-length genotype 2a HCV genome (JFH-1) that replicates and produces infectious virus in cell culture,2 the entire HCV life cycle is now accessible for investigation. To date, HCV infection and replication in vitro have been studied in actively dividing poorly differentiated human hepatoma Huh-7 cells and derivatives. Therefore, despite its strong advantages, the cell-culture grown HCV (HCVcc) system cannot completely mimic the events which occur during a natural HCV infection in vivo. To improve this system, Huh-7 cells were either cultured in the presence of DMSO to induce hepatocyte-specific genes, or in a 3D environment to promote cell polarization.3 Immortalized human hepatocytes with impaired interferon (IFN) signaling or in a 3D culture system (HuS-E/2)4 were also shown to be useful tools for the study of HCV infection. However, the production of HCV particles in these cells appeared to be restricted to a short period. Freshly isolated primary human hepatocytes (PHHs) are obviously the most exquisite system to study HCV infectivity. However they are scarce, with unpredictable availability and phenotypically unstable, have limited growth potential and life span, and exhibit large interdonor variability. Moreover, when serum-derived HCV particles were used to infect PHHs, low-efficient short-time virus production was observed. Even if PHHs were shown to be sensitive to HCVcc, the form of virus particle and the host cell are important determinants for virus entry and HCV replication.4

In this context, HepaRG cells offer a unique opportunity. These are indeed progenitor cells capable of differentiating into hepatocytes and biliary epithelial cells depending on culture conditions.5 HepaRG hepatocytes have bile canalicular-like structures and present well-localized zonula occludens protein-1 (ZO-1) as a tight junction-associated protein.5 The HepaRG cells stably express for a long time (over up to a 1-month confluence period) various liver-specific functions, including the major cytochromes P450 (CYP1A2, 2B6, 3A4, and 2E1), and exhibit a similar gene expression pattern as PPHs and human liver tissues.6 Thus, HepaRG cells could provide an attractive alternative to PHHs and represent a promising cellular model to study virus-host interactions. Thanks to a unique monoclonal antibody (mAb) D32.10, we have previously succeeded in characterizing native circulating enveloped HCV particles, designated HCVsp, in chronic hepatitis C patients.7, 8 Remarkably, mAb D32.10 has been shown to be so far the only antibody able to efficiently inhibit the interactions between HCVsp and hepatocytes.9

The aim of this study was therefore to investigate whether progenitors and/or differentiated HepaRG cells could be directly infected with HCVsp and sustainably propagate HCV RNA-containing enveloped particles to further assess the D32.10 mAb neutralizing properties and screen for similar antibody activity.


apoB, apolipoprotein B; apoE, apolipoprotein E; ELISA, enzyme-linked immunosorbent assay; EM, electron microscopy; ER, endoplasmic reticulum; GAG, glycosaminoglycan; HDL, high-density lipoprotein; HCV, hepatitis C virus; HCVcc, infectious cell culture HCV particle; HCVpp, HCV pseudotyped particle; HCVsp, serum-derived HCV particle; HCV-RG, HCV-infected HepaRG cells; IC50, the inhibitory antibody concentration reducing the binding by 50%; IEM, immune-EM; IgG, immunoglobulin G; IHC, immunohistochemistry; LDL-R, low density lipoprotein-receptor; mAb, monoclonal antibody; m.o.i., multiplicity of infection; MVB, multivesicular bodies; NHS, normal human serum; PHH, primary human hepatocytes; RT-qPCR, reverse transcription followed by quantitative real-time polymerase chain reaction.

Materials and Methods

Cell Culture and HCVsp-HepaRG Infection Experimental Protocol.

HepaRG cells were cultured as described.5, 9 The medium was renewed every 7 days. HCV infection experiments were carried out using well-characterized purified HCVsp (genotype 3a), which contained 106 copies of HCV RNA per mg of protein (Supporting Materials and Methods, Fig. 1). HepaRG cells were inoculated 3 days postplating (p.p.) for 18 hours at 37°C with 105 copies of HCV RNA corresponding to multiplicity of infection (m.o.i.) of 1, either in the absence or in the presence of normal human serum (NHS) at a final concentration of 1%. For inhibition experiments, the HCVsp inoculum was preincubated for 2 hours at 37°C with the D32.10 mAb at a final concentration of 0.5 μg/mL. On day 1 postinfection (p.i.), extensive washings (3-4 times) of the cells were done. The medium was changed at day 7 p.p. and then each week at days 14, 21, 28 up to day 66. HCV-associated particles were purified from supernatants collected every 7 days and clarified by low-speed centrifugation, as described.10 Cells were harvested at days 28 and 56 p.p. for detection of E1E2 and core antigens by immunohistochemistry. HCVsp-infected HepaRG cells were frozen at day 56 p.p.

Buoyant Density Determination in Sucrose or Iodixanol Gradients.

The density distributions of secreted HCV particles were analyzed on either a sucrose density gradient (10%-60% w/w)10 or an iodixanol (OptiPrep; Axis-Shield, AbCys S.A. France) gradient prepared as described by Nielsen et al.11 HCV particles concentrated (50-100 times) and purified from culture media were subjected to isopycnic centrifugation (200,000g for 48 hours at 4°C) in the SW41 rotor of a Beckman centrifuge. Fractions (0.6 mL or 1.2 mL) were collected from the bottom of the tube and the density of each was determined by refractometry.

Detection of HCV RNA (+) by One-Step Reverse-Transcription Polymerase Chain Reaction (RT-PCR) and Quantification by Real-Time PCR.

For HCV RNA analysis, lysis buffer was directly added to the ultracentrifugation pellets. RNA was extracted using QIAamp Viral RNA mini Kit (Qiagen). Reverse transcription was performed using primers located in the 5′ NCR region of all HCV genotypes.12 After a denaturing step, the RNA template was incubated at 60°C for 1 hour with 7.5 U thermoscript reverse transcriptase (kit Gibco/BRL) and then treated with 20 U RNaseOut for 20 minutes at 37°C. Quantitative PCR (qPCR) was carried out using EXPRESS One-step SYBR GreenER qRT-PCR (Invitrogen) with iCycler Bio-Rad instruments and technology. Primers and PCR protocols are detailed in Supporting Materials and Methods. Quantitation was performed using an internal standard curve (JFH-1 RNA: 1 pg to 10 ng).

E1E2 Antigen Titration and Apolipoprotein E/Apolipoprotein B (apoE/apoB) Association by In-House Indirect Enzyme-Linked Immunosorbent Assay (ELISA).

Wells were coated with purified, concentrated virus particles from 0.1 to 40 μg of protein/mL or gradient fractions at dilutions 1/2, 1/10, or 1/100. E1E2 antigenic activity was analyzed as described.7, 8, 10 apoE and apoB association with virus particles was determined by indirect ELISA using goat polyclonal antibodies to apoE (ab7620) or to apoB 40/100 (ab27626) from Abcam (Paris, France) as primary antibodies and antigoat specific antibody horseradish peroxidase (HRP) conjugate as secondary antibodies. The results were considered positive (P) when superior to the cutoff, corresponding to the mean of negative (N) controls multiplied by 2.1, i.e., P/N ratio >2.1.

Quantification of Core Antigen.

HCV core antigen levels in purified, concentrated virus particles or gradient fractions (dilutions 1/2, 1/10) were quantified by a two-step ELISA system using the Ortho HCV antigen ELISA test kit from Wako Chemicals (Neuss, Germany). The results were considered positive when >50 fmol/L.

Immunohistochemistry (IHC).

HepaRG cells grown on slides were fixed with 2% paraformaldehyde for 30 minutes at room temperature and washed 3 times in phosphate-buffered saline (PBS). The immunostaining was performed using the R.T.U. Vectastain Universal Elite ABC kit from Vector laboratories (AbCys S.A. France) with primary antibodies to HCV E1E2 (D32.10 mAb) or core (C7-50 mAb) proteins, as detailed in Supporting Materials and Methods and Table 1.

Electron Microscopy (EM) and Immune-EM Analyses.

For ultrastructural analysis by EM, cells (≥106) were fixed for 30 minutes at room temperature with 4% glutaraldehyde in culture medium (50:50, v/v), and then with 4% glutaraldehyde in 0.2 M cacodylate buffer, pH 7.4. After postfixation and dehydration steps, cell pellets were embedded in Epon resin (see Supporting Materials and Methods). For observation, ultrathin sections (60-70 nm thick) were cut, deposited on copper grids, and stained with 1% uranyl acetate-1% lead citrate. For intracellular localization of viral and cellular proteins, cells were fixed for 1 hour at room temperature followed by 1 hour at 4°C with 2% PLP metaperiodate in 0.1 M phosphate buffer (pH 7.2). Cell pellets were embedded in LR White resin. Ultrathin sections were deposited on nickel grids for immunogold labeling (see Supporting Materials and Methods and Table 1). Primary antibodies used were the anti-E1E2 D32.10 mAb, or a polyclonal anti-HSC70 goat antibody. The grids were incubated with a 1/80 dilution of secondary gold-conjugated goat antimouse (gold beads of 10 nm or 20 nm in diameter) or rabbit antigoat (gold beads of 5 nm in diameter) from BioCell Research Laboratories, then stained as described above. Grids were examined using a JEM Jeol 1400 electron microscope (JEOL, Tokyo, Japan) equipped with a Gatan Orius 600 camera driven by Digital Micrograph logical.


HepaRG Progenitor Cells Are Susceptible to HCVsp Infection and Differentiated HepaRG Hepatocytes Produce Infectious Enveloped HCV Particles.

Because it has been shown that serum-dependent mechanisms facilitated the entry of HCVpp and HCVcc into target cells,13 HCVsp infection was performed at day 3 p.p. either in the absence (Fig. 1A,a) or in the presence (Fig. 1A,b) of 1% NHS only during the infection process. Culture medium was subjected to ultracentrifugation for removing cellular material and proteins and pelleting HCV RNA-associated particles. As illustrated in Fig. 1A, HCV RNA(+) was detected by RT-PCR at days 7 and 14 p.p. in the absence of NHS (a) or at days 4 and 7 p.p. in the presence of NHS (b), reflecting adsorption and/or partial penetration of the inoculum into the cells. No detection was observed at day 21 (−NHS) or 14 (+NHS) p.p. Thus, residual cell-surface bound HCV RNA was completely eliminated (<100 copies/mL, which is the detection limit of the qPCR technique used). Thereafter, the extracellular HCV RNA progressively increased and reached 6log10 copies/mL at days 42-49 (+NHS) corresponding to production of newly synthesized HCV RNA-positive particles. The infection was efficient because the HCV RNA increased by 4 logs from 14 to 49 days (+NHS). These results reflect that HCV actively replicated in the infected cells and spread into uninfected cells because the HepaRG cells did not proliferate during the differentiated phase of the culture. Interestingly, whereas a common cyclic pattern was observed when the infection was performed without NHS (a), a continuous pattern occurred in the presence of 1% NHS (b). More rapid penetration of HCVsp and best synchronization of infection resulted from the presence of NHS during the infection. This condition was therefore used in subsequent experiments.

Figure 1.

Production of infectious enveloped HCV particles by differentiated HepaRG hepatocytes. Clarified supernatants harvested at different times were subjected to ultracentrifugation for concentration and purification of HCV-related particles. (A) Detection and quantification of HCV RNA levels in the absence (a) or in the presence of 1% NHS (b). As positive control (pos), the inoculum HCVsp was loaded in the last lane. (B) Kinetics of E1E2 antigen (Ag) by indirect ELISA (a) and western blot (WB, b) using the D32.10 mAb. Purified, concentrated virus particles were titrated. The cutoff for 1/10 and 1/100 dilutions was visualized and corresponded to P/N ratio = 2.1. HCVcc was loaded in the last lane as positive control in WB (Supporting Materials and Methods). (C) Purified, concentrated HCV particles from D28* (Fig. 1A,b) were used to reinfect naïve HepaRG cells. HCV RNA(+) was detected by RT-PCR in HCV particles pelleted from supernatants collected from D4 to D28 p.p. (D) Pooled particles from D28+D42 were subjected to isopycnic iodixanol gradient ultracentrifugation. All the fractions were probed for E1E2 Ag (P/N ratio at 1/100 dilution), the fractions 3, 4, and 6 for RNA levels in RT-qPCR, the fraction 5 for core Ag in commercial ELISA, and the fractions 3, 4, and 5 for apoE and apoB in indirect ELISA. (E) HCV E1E2 and core proteins detection by IHC. IHC was performed in control HCV(−) uninfected HepaRG cells and in HCVsp-infected HepaRG cells at D28*. (a): Anti-E1E2/D32.10 mAb (5 μg/mL). (b): Anti-HCV core mAb (clone C7-50, Abcam) (1 μg/mL).

The HCV amplification was also assessed by determination of E1E2 antigenic activity (Fig. 1B) by indirect ELISA (a) and western blotting (b) using the D32.10 mAb. The increased HCV E1E2 in the medium from day 28 to day 49 p.p. correlated well with the HCV RNA peak detection and supported de novo synthesis and release of enveloped RNA-containing particles. The cutoff values were calculated by using three control samples from uninfected HepaRG cells (mean optical density values: 0.296 ± 0.124 for 1/10 dilution, 0.093 ± 0.025 for 1/50 dilution, and 0.071 ± 0.010 for 1/100 dilution). A good correlation between ELISA and western blot techniques was observed. Examination of HCV core antigen with a commercial immunoassay confirmed the production of complete virions containing both HCV RNA and core antigen and expressing E1E2 envelope proteins.

Next we asked whether the HCV particles produced by HCVsp-infected HepaRG cells were infectious. To this end, the ability of HCV particles released into the cell culture media 25 days after infection 1 (corresponding to D28* p.p., cf. Fig. 1A,b) to infect naive HepaRG cells at 3 days p.p. was tested (Fig. 1C). HCV RNA(+) was analyzed by RT-PCR in the supernatants collected each week and subjected to ultracentrifugation as described above. After early detection at day 1 p.i., extracellular HCV RNA strongly decreased during the proliferating (“P”) phase and HCV RNA amplification was then observed during the differentiation (“D”) phase. This indicates that the HCV particles released from infected-HepaRG cells (HCV-RG) are indeed infectious.

To determine the buoyant density distribution of HCV RNA, E1E2 and core antigens, the viral preparations from media collected at days 28 and 42 p.p. (Fig. 1A,b) were pooled and subjected to iodixanol gradient density centrifugation. Figure 1D shows that the HCV-RG particles had a relatively homogeneous distribution between 1.06 and 1.12 g/mL. They expressed E1E2 envelope proteins and contained RNA and core antigen. In addition, the positive fractions reacted with polyclonal antibodies against apoE (++, P/N ratio = 5-6) and apoB (+, P/N ratio = 3-4), suggesting that host lipoproteins could be associated with these particles mimicking circulating HCV.14

Immunohistochemistry experiments were performed to investigate intracellular expression of HCV E1E2 and core antigens (Ag) in infected-HepaRG cells at 28 and 56 days p.p. (infection 1). Figure 1E shows that the HCVsp-infected HepaRG cells at D28 p.p. exhibited a very strong staining of cytoplasm and perinuclear regions for E1E2 Ag (a). Fifty to sixty percent of cells were positive. Core Ag staining (b) appeared also in the cytoplasm possibly around lipid droplets. Some cells were labeled both in the cytosol and the nucleus. Control HCV(−) uninfected HepaRG cells were clearly negative in the presence of D32.10 (a) or C7.50 (b), as well as HCV-infected cells in the presence of a control IgG1 antibody (not shown). Positively stained infected cells exhibited morphological features of hepatocytes.5

Altogether, these results indicate that the human HepaRG cells can be infected with HCVsp when proliferated and do produce de novo infectious lipoprotein-associated enveloped complete HCV particles for up to 6 weeks when differentiated.

HCVsp Infection of HepaRG Cells Is Inhibited by the Anti-HCV E1E2 D32.10 mAb.

To investigate whether the unique E1E2-specific D32.10 mAb inhibits HCV infection, the infection experiment (infection 3) was performed after preincubation of HCVsp with D32.10 at a 0.5 μg/mL concentration. Figure 2 shows that the D32.10 mAb completely inhibited HCV RNA production in HepaRG culture supernatants. The total amount of HCV RNA remained at very low levels throughout the follow-up of the infection from day 1 to day 21 in the presence of D32.10 with a mean inhibition of 80.5 ± 11.6% (Fig. 2A). When HCV RNA was quantified by qPCR, 5log10 copies/mL were detected at day 21 after control infection. The preincubation of the inoculum with D32.10 reduced by ≈97% the extracellular HCV RNA (−2 log10, Fig. 2B). To further support that control-infected HepaRG cells produced viral particles, iodixanol density gradient analysis was performed from HCV RNA-associated particles present in the culture media collected at days 14 and 21 (Fig. 2C). As seen previously (infection 1), both HCV RNA, E1E2, and core antigens were recovered as a major peak between 1.08 and 1.15 g/mL, reflecting the reproducibility of our infection system. No particles could be detected when HCV infection was performed in the presence of D32.10 (all the values were under the cutoffs, not shown). These results indicate that the D32.10 mAb efficiently inhibits HCVsp infection of HepaRG hepatocytes.

Figure 2.

Inhibition of HCVsp infection in HepaRG cells by the anti-E1E2 D32.10 mAb. HCVsp inoculum was preincubated for 2 hours at 37°C with the D32.10 mAb at a concentration of 0.5 μg/mL (or control IgG1) and HepaRG cells were inoculated at day 3 p.p. (A) Kinetics of HCV RNA levels detected by RT-PCR in HepaRG culture supernatants. (B) Quantification of extracellular HCV RNA by RT-qPCR at days 14 and 21 p.i. (C) Pooled particles from D14+D21 were subjected to isopycnic iodixanol gradient ultracentrifugation. Only control infection samples (HCVsp) gave positive results. All the fractions were probed for E1E2 Ag (♦, P/N ratio at ½ dilution; ▪, P/N ratio at 1/10 dilution), the fractions 2, 4, 6, and 7 for RNA levels in RT-qPCR, the fractions 4, 6, and 7 for core Ag in commercial ELISA.

HepaRG Cells Can Become Persistently Infected by HCVsp (HCVsp-RG Cells).

To assess whether the differentiated HepaRG cells could indeed become persistently infected with HCVsp, cells were frozen at D56** p.p. after primary infection. The HCV-infected HepaRG cells were then thawed, plated at low density (4 × 104/cm2), and subjected either to 1 (P1) or 3 (P3) subcultures before forcing induction of the differentiation process (Fig. 3A). The supernatants were then collected each week and analyzed as above. Figure 3B shows that extracellular HCV RNA could be detected only during the differentiation (“D”) stage between 14 and 28 (P1) or 35 to 56 (P3) days. Interestingly, earlier (D14 instead of D35) and higher (4.5 log10 instead of 3.3 log10 copies/mL) titer virus levels were observed after one rather than three subcultures. This suggests that successive phases of proliferation (“P”) before induction of the differentiation process (no splitting at confluency) resulted in an advantage of noninfected over HCV-infected HepaRG cells. When we analyzed the HCV particles on sucrose gradient released in the culture media collected at D28 (P1) and D49 (P3) as a pool corresponding to 4.7 log10 copies of HCV RNA/mL, the total amount of HCV RNA cosedimented with core antigen and E1E2 in association with apoE and apoB at densities between 1.18 and 1.20 g/mL (peak II, Fig. 3C). In these experimental conditions of fractionation,7, 10 no reactivity was detected at low density (1.06 g/mL). However, a major peak of defective particles containing only E1E2 envelope associated with lipoproteins (apoE and apoB) sedimented at intermediate densities of 1.14-1.15 g/mL (peak I, Fig. 3C).14

Figure 3.

Persistent infection of HepaRG cells with HCVsp. (A) HCVsp-infected HepaRG cells (Inf. 1) were frozen at day 56 p.p. and reseeded (Plating) at low density (4 × 104/cm2). The differentiation (“D”) process was induced (no splitting) after one (P1) or three (P3) successive proliferative phases (D0-D7) by subculturing at day 7 postplating. (B) Extracellular HCV RNA was quantified by RT-qPCR at different times. HCV core Ag was assayed in pooled samples from D14+D28 (Inf. 2P1). (C) Pooled particles from D28 (Inf. 2P1) and D49 (Inf.2P3) were subjected to isopycnic sucrose gradient ultracentrifugation. All the fractions were probed for E1E2 Ag (♦, P/N ratio at 1/10 dilution; ▪, P/N ratio at 1/100 dilution). The fractions 5, 11, 12, and 14, 15, 16 were probed for RNA levels in RT-qPCR, core Ag in commercial ELISA, and apoE and apoB in indirect ELISA. P/N ratio = 2-3, +lim; P/N ratio = 3-4, +; P/N ratio = 5-6, ++; P/N ratio > 6, +++.

Taken together, these data demonstrate that the HepaRG cells remained persistently infected by HCVsp (HCVsp-RG cells) and could produce larger amounts of empty apoE/apoB-associated E1E2 than apoE/apoB-associated complete HCV particles only when still differentiated.

Ultrastructural Aspects of HCVsp-RG Cells.

To identify ultrastructural modifications induced by HCVsp infection, EM analysis was performed. The HCV-infected HepaRG cells frozen at day 56** after plating (infection 1: Fig. 1A,b) were thawed, seeded at low density, cultivated 1 week until confluence, reseeded (P1, Fig. 3A), and then maintained without splitting after confluency up to day 28. At this time, EM examination of noninfected HepaRG cells revealed characteristics typical of normal human hepatocytes (Fig. 4A). Apical and basolateral poles as well as tight junctions between two adjacent hepatocyte-like cells (H) were clearly identified (Fig. 4A). Bile canalicular structures with microvilli protruding into the lumen and no alterations in the endoplasmic reticulum (ER) were observed. This confirms that HepaRG hepatocytes have membrane polarity and are well differentiated, closely resembling their counterparts in vivo.5, 6 EM of HCVsp-RG cells revealed the presence of several morphological alterations (Fig. 4B). A membranous web composed of small vesicles was identified in many cells (a). Multiple small vesicles connected to the membrane were observed (b). Multivesicular bodies (MVBs) accumulated internal vesicles (c). Submembranous thickening of cytoskeleton at the apical pole was visualized (d). Remarkably, karmellae-like, multilayer structures typical of membrane rearrangements associated with RNA replication by varied (+)RNA viruses15 were found specifically in HCVsp-RG cells (e). Some rare HCVsp-RG cells exhibited typical apoptosis-associated morphological alterations like formation of apoptotic bleeds (f).

Figure 4.

EM analysis of HCVsp persistently infected HepaRG cells (HCVsp-RG). (A) Ultrastructural aspects of HCV(−) noninfected HepaRG cells. H, hepatocyte-like cells. Nuc, nucleus. np, nuclear pores. Nucl, nucleoli. ER, endoplasmic reticulum. (B) Main ultrastructural changes in HCVsp-RG cells (see Results).

HCV E1E2 Intracellular Localization in HCVsp-RG Cells.

Immuno-EM was performed to localize E1E2 Ag recognized by D32.10 in HCVsp-RG cells at the same culture time as morphological studies. Figure 5A shows that immunogold labeling for E1E2 was observed associated with 40-100 nm vesicles budding at the plasma membrane, resembling exosomes. To support potential association of E1E2 with exosomes, double-label immunogold EM experiments were performed using anti-E1E2/D32.10 (20 nm) and anti-HSC70 (5 nm). As shown in Fig. 5B, colabeling of HSC70 with E1E2 on the internal vesicles accumulated under the plasma membrane was observed. No immunolabeling with D32.10 was detectable in the noninfected HepaRG control cells (data not shown).

Figure 5.

Immune-EM analysis of HCVsp-RG cells. Cryosections were labeled using either anti-E1E2/D32.10 mAb alone (A) or in combination with the anti-HSC70 (K19) polyclonal antibody (B). Antibodies bound were detected using antimouse or antigoat secondary antibodies conjugated to 20 nm and 5 nm gold particles, respectively. Cyt, cytosol.


The differentiation-inducible properties and the typical features of fully functional mature hepatocytes exhibited by HepaRG cells5, 6 make them attractive candidates for infection with naturally circulating HCV particles isolated from chronically infected patients.7 Interestingly, the infection was primed in progenitors, whereas relatively robust sustainable replication and propagation of the infection only occurred in fully differentiated HepaRG cells with HCV RNA amplification up to 6 log10 for at least 1 to 2 months. Remarkably, the presence of 1% NHS during the infection process of HepaRG cells with HCVsp resulted in a more rapid internalization and steady HCV RNA production in culture supernatants from 3 up to 9 weeks. This supports a possible synchronization of infection through serum factors such as high-density lipoproteins (HDLs), which have been shown to facilitate the entry of HCVpp and HCVcc into target cells.13 The endocytosis of viral particles could thus be accelerated by suppression of a time lag in which cell-bound virions are not internalized. These conditions appeared therefore optimal for mimicking natural infection.

Because of the weak, very transient, delayed, and often artifactual detection of negative-strand viral RNA in infected cells, the HCV RNA amplification in the culture medium as enveloped complete virions10 and the detection of HCV structural proteins in the cells were used as infectivity assays. HCV infection was validated by different approaches: (1) time-dependent amplification of newly synthesized HCV RNA and E1E2 proteins as particulate forms in media; (2) production of E1E2/core/RNA(+)-particles which sedimented at 1.08 g/mL in iodixanol gradient (corresponding to 1.18 g/mL in sucrose gradient) and were infectious as indicated by passage to naïve HepaRG cells; (3) HCV E1E2 and core protein accumulation in the cytoplasm of infected cells 1 month p.i.; (4) complete reduction of HCV RNA and infectious virus in HepaRG culture supernatants (97% at 3 weeks p.i.) by E1E2-specific mAb D32.107-9 at low concentration (0.5 μg/mL) even when the infection was performed in the presence of NHS; (5) ability of infected-HepaRG cells to produce high titers of HCV RNA (4 to 5log10) as complete virus particles in culture media after freezing/thawing and subculture(s) followed by induction of the differentiation process; (6) production of apoE/apoB-associated HCV virions by the HepaRG cells similar to authentic patient-derived HCV particles11, 14; (7) observation of typical positive-strand RNA virus-induced membrane rearrangements18 and detection of HCV E1E2 antigen in association with vesicular structures in ER and at a submembranous localization in HCVsp-RG cells.

Very recently, a cell-culture-based system was established using PHHs inoculated with HCVcc.16 Even if freshly isolated PHHs are currently the in vitro “gold standard” of human liver cells, the HepaRG human hepatic cell line is now increasingly used as a surrogate for PHHs in pharmaceutical research and development for metabolism studies.17 Here, our results demonstrate that HepaRG cells can be infected with serum-derived HCV of genotype 3 and persistently produce infectious enveloped HCV particles with biophysical and immunological properties similar to circulating7, 11 and infected liver-derived10 HCV. The major contributions of our study were to use a genuine HCV isolate from patients distinct from the JFH-1 or Jc1 virus of genotype 2a together with the HepaRG cell line, which possesses key features of authentic hepatocytes. Of course, the current Huh-7-derived HCVcc system remains the “gold standard,” and it would have been optimal to successfully infect HepaRG cells with HCVcc. Unfortunately, only a weak transient replication was obtained in our laboratory when we tried to inoculate differentiated HepaRG cells with a highly infectious JFH-1 inoculum (Durantel et al., unpubl. data). This could be due to the production of type-I interferons in the culture medium,18 which likely should inhibit HCV replication and spreading. This could also explain why the HepaRG cells are only susceptible to HCVsp infection when they exhibit dedifferentiated, depolarized epithelial phenotype associated with an immature innate immunity, resistance to apoptosis, and cellular growth.19 Moreover, the mature hepatocyte phenotype associated with the appearance of liver-specific functions after growth arrest6 seems to be essential for the ability of HepaRG cells to further replicate and propagate HCVsp efficiently. Indeed, different experimental protocols of infection were initially performed: at the proliferative or differentiated stage of culture, with addition or not of NHS during the infection process, and of 2% dimethyl sulfoxide (DMSO) to the culture medium to force the differentiation process. The best conditions were HCV infection at the proliferative stage (day 3 p.p.) in the presence of 1% NHS and absence of DMSO in the differentiation medium.

To further validate our HCV infection system, EM and immune-EM analyses of HCVsp-RG cells were performed at the differentiated stage when cells produced HCV particles. Typical ultrastructural changes were visualized, resembling those observed in hepatocytes of chronically HCV-infected patients,20 and found associated with JFH-1 strain replication.21 The biogenesis of exosomes from the endosomal system as powerful intercellular messengers differs in polarized and nonpolarized cells.22 Therefore, the export of HCV particles with formation of virus-containing small vesicles that resemble exosomes, like those of other enveloped RNA viruses, may be specifically associated with the hepatocytic differentiation status of HepaRG cells. Colabeling of E1E2 and HSC70, a chaperone protein identified in exosomes22 and as an HCV virion-associated protein,23 could support an association of HCV envelope proteins with exosomes through CD81 for release into the extracellular milieu.24

Finally, the HCVsp-HepaRG infection system may be used to test cell entry “blockers.” Here, as a preliminary result, we demonstrated that the infection could be efficiently inhibited by pretreatment of the virus with the unique D32.10 mAb. This supports that the binding of D32.10 to its E1E2-specific discontinuous antigenic determinant on HCVsp7 may directly block the first steps of virus entry into HepaRG cells. Indeed, the regions in the E2 glycoprotein recognized by D32.10 contain glycosaminoglycan (GAG)- and CD81-binding sites. By using CD81 antibody for blocking HCVsp binding to HepaRG cells, as described,9 a dose-dependent inhibitory effect was observed with an IC50 of 1 μg/mL (Supporting Results and Fig. 2). Our studies in vivo in HCV-infected patients showed that anti-E1E2 D32.10 epitope-binding antibodies were strictly generated from patients who cleared HCV either spontaneously or after achieving a sustained viral response to antiviral therapy.26 Convergence of in vitro and in vivo data supports the virus-neutralizing activity of the D32.10 mAb, and the targeting of the D32.10 epitope by host neutralizing responses during HCV infection.

In conclusion, our results show that, whereas hepatic progenitors can be infected with naturally occurring HCVsp of genotype 3, only the fully differentiated HepaRG hepatocytes can produce infectious apoE/apoB-associated enveloped HCV particles. The early complete inhibition of primary infection of HepaRG cells with HCVsp by the D32.10 mAb in the presence of NHS may favor the efficient neutralizing activity through a direct association with the E1E2 envelope glycoproteins. The unique composite E1E2/D32.10 epitope seems to be essential for HCVsp entry and thus the D32.10 mAb a novel inhibitor of HCV infection, with most relevant potential in the context of liver transplantation to prevent reinfection of the graft. This new HCVsp-HepaRG infection system could also be most useful for screening HCV entry inhibitors.