Characterization of host-range and cell entry properties of the major genotypes and subtypes of hepatitis C virus


  • Dimitri Lavillette,

    1. Laboratoire de Vectorologie Rétrovirale et Thérapie Génique, INSERM U412, IFR128, BioSciences Lyon-Gerland, Ecole Normale Supérieure de Lyon, Lyon, France
    Search for more papers by this author
    • D.L. and A.W.T. contributed equally to this work.

  • Alexander W. Tarr,

    1. Institute of Infection, Immunity and Inflammation and Division of Microbiology, The University of Nottingham, Queen's Medical Centre, Nottingham, United Kingdom
    Search for more papers by this author
    • D.L. and A.W.T. contributed equally to this work.

  • Cécile Voisset,

    1. CNRS-UPR2511, Institut de Biologie de Lille-Institut Pasteur de Lille, Lille, France
    Search for more papers by this author
  • Peggy Donot,

    1. Laboratoire de Vectorologie Rétrovirale et Thérapie Génique, INSERM U412, IFR128, BioSciences Lyon-Gerland, Ecole Normale Supérieure de Lyon, Lyon, France
    Search for more papers by this author
  • Birke Bartosch,

    1. Laboratoire de Vectorologie Rétrovirale et Thérapie Génique, INSERM U412, IFR128, BioSciences Lyon-Gerland, Ecole Normale Supérieure de Lyon, Lyon, France
    Search for more papers by this author
  • Christine Bain,

    1. FRE 2736 CNRS/bioMérieux, IFR128 BioSciences Lyon-Gerland, Lyon, France
    Search for more papers by this author
  • Arvind H. Patel,

    1. MRC Virology Unit, Institute of Virology, University of Glasgow, Glasgow, United Kingdom
    Search for more papers by this author
  • Jean Dubuisson,

    1. CNRS-UPR2511, Institut de Biologie de Lille-Institut Pasteur de Lille, Lille, France
    Search for more papers by this author
  • Jonathan K. Ball,

    1. Institute of Infection, Immunity and Inflammation and Division of Microbiology, The University of Nottingham, Queen's Medical Centre, Nottingham, United Kingdom
    Search for more papers by this author
  • François-Loïc Cosset

    Corresponding author
    1. Laboratoire de Vectorologie Rétrovirale et Thérapie Génique, INSERM U412, IFR128, BioSciences Lyon-Gerland, Ecole Normale Supérieure de Lyon, Lyon, France
    • LVRTG. ENS de Lyon, 46 Allée d'Italie, 69364 Lyon Cedex 07, France
    Search for more papers by this author
    • fax: (33) 472 72 80 80

  • Conflict of interest: Nothing to report.


Because of the lack of a robust cell culture system, relatively little is known about the molecular details of the cell entry mechanism for hepatitis C virus (HCV). Recently, we described infectious HCV pseudo-particles (HCVpp) that were generated by incorporating unmodified HCV E1E2 glycoproteins into the membrane of retroviral core particles. These initial studies, performed with E1E2 glycoproteins of genotype 1, noted that HCVpp closely mimic the cell entry and neutralization properties of parental HCV. Because sequence variations in E1 and E2 may account for differences in tropism, replication properties, neutralization, and response to treatment in patients infected with different genotypes, we investigated the functional properties of HCV envelope glycoproteins from different genotypes/subtypes. Our studies indicate that hepatocytes were preferential targets of infection in vitro, although HCV replication in extrahepatic sites has been reported in vivo. Receptor competition assays using antibodies against the CD81 ectodomain as well as ectopic expression of CD81 in CD81-deficient HepG2 cells indicated that CD81 is used by all the different genotypes/subtypes analyzed to enter the cells. However, by silencing RNA (siRNA) interference assays, our results show that the level of Scavenger Receptor Class-B Type-I (SR-BI) needed for efficient infection varies between genotypes and subtypes. Finally, sera from chronic HCV carriers were found to exhibit broadly reactive activities that inhibited HCVpp cell entry, but failed to neutralize all the different genotypes. In conclusion, we characterize common steps in the cell entry pathways of the major HCV genotypes that should provide clues for the development of cell entry inhibitors and vaccines. (HEPATOLOGY 2005;41:265–274.)

Hepatitis C virus (HCV) is an RNA-enveloped virus belonging to the Flaviviridae family. HCV exhibits a high degree of genetic heterogeneity. The propensity for genetic change is associated primarily with the error-prone nature of its RNA-dependent RNA polymerase together with the high HCV replicative rate in vivo.1, 2 This results in infected individuals harboring a diverse population of viral variants known as a quasispecies, which evolve in response to a variety of selective pressures.3 Although HCV-specific immunity develops after primary infection, it frequently fails to eliminate the virus.4–6 HCV has infected approximately 170 million people worldwide, and approximately 80% of those infected will develop chronic infection. The outcome of chronic infection varies widely between individuals, but large proportions of patients develop serious liver diseases such as cirrhosis and hepatocellular carcinoma.7 Current therapies are inadequate, and development of appropriate therapeutic and prophylactic vaccines remains a significant challenge.

HCV can be classified into six genetically distinct genotypes and further subdivided into at least 70 subtypes, which differ by approximately 30% and 15% at the nucleotide level, respectively.8 The different genotypes may exhibit differing phenotypic properties. For example, infection by genotype 1 viruses may be associated with development of more severe liver disease9 and more limited response to antiviral therapy.10 Furthermore, high levels of variability are seen particularly in the envelope genes, and such heterogeneity is likely to pose a significant challenge for vaccine design.11

Virus entry is mediated by the viral envelope glycoproteins E1 and E2. These proteins are thought to be anchored in the viral lipid membrane at the surface of the virus, as noncovalent heterodimers12–14 extensively modified by N-linked glycosylation.15 Investigation of the binding and entry of virions has been hampered by the inability to obtain large quantities of virions from patient samples16 and the difficulties in amplifying the virus in vitro.17 Despite this, studies using cell binding assays together with infection assays using HCV pseudo-particles (HCVpp) have shown that HCV entry into isolated primary liver cells and cell lines requires interaction with the cell surface receptors CD81 and Scavenger Receptor Class-B Type-I (SR-BI),13, 18–24 at least for those viruses belonging to genotype 1. However, the presence of these receptors alone is not sufficient to allow viral particle entry, indicating that additional factors or cell receptors are required to facilitate entry. A number of recent investigations have suggested that glycoproteins from different genotypes might function differently during the entry pathway. For example, studies with a soluble form of the E2 protein have highlighted genotype-specific differences in CD81 binding affinity,20 even to the extent that genotype 3 of HCV was unable to bind to CD81.25

Thus, to investigate potential differences in the entry pathways used by HCV belonging to different genotypes, here we have generated a panel of full-length HCV glycoprotein clones, representative of genotypes 1 through 6, and assessed whether they were capable of allowing entry in the HCVpp assay. We found that all genotypes of HCV are hepatotropic, with an absolute requirement for CD81 during infection. In contrast, the role of SR-BI in viral entry appeared to vary depending on genotype. Finally, sera from chronic HCV carriers were found to exhibit broadly reactive activities that inhibited HCVpp cell entry, but failed to neutralize all of the different genotypes. Altogether, these findings have implications for the development of cell entry inhibitors and vaccines.


HCV, hepatits C virus; HCVpp, HCV pseudo-particles; SR-BI, Scavenger Receptor Class-B Type-I; MLV, murine leukemia virus; TU, transducing units; siRNA, silencing RNA; HIV, human immunodeficiency virus; sE2, soluble forms of the E2 protein.

Materials and Methods

Cell Lines.

293T (ATCC CRL-1573), Huh-7, PLC/PRF/5 (CRL-8024), Hep3B human hepatocellular carcinoma (HB-8064), HepG2 (HB-8065), HepG2-CD81,22 SW-13 (CCL-105), 293T SR-BI (22), HOS (CRL-1543), Hela (CCL-2), TE671 (CRL-8805), Molt-4 (CRL-1582), Jurkat (TIB-152), Raji (CCL-86) and Akata, THP-1, U87 (HTB-14), U118 (HTB-15), A431 (CRL-1555), COS-7 (CRL-1651), Vero (CCL-81), CHO (CCL-61) and CHO CD81/SR-BI22 were grown as recommended by the ATCC (American Type Culture Collection, Rockville, MD). Huh-7 and Hela cells were transduced with a lentiviral vector coding for L-SIGN (gift from Pierre-Yves Lozach and Ralf Altmeyer Institut Pasteur, Paris, France). THP cells expressing L-SIGN and THP DC-SIGN (gift from Pierre-Yves Lozach and Ralf Altmeyer) were grown in RPMI-10% fetal calf serum.

Isolation of E1 and E2 Glycoproteins of Different Genotypes and Expression Constructs.

DNA sequences of envelope E1 and E2 glycoproteins were isolated from patients infected with different genotypes of HCV. Viral RNA was isolated from 100 μL serum using a total RNA extraction kit (Fluka). Virus-specific cDNA was generated by using genotype-defined primers (Table 1) with a First Strand cDNA Kit (Amersham). From 1 μL of cDNA, polymerase chain reaction products were generated representing amino acid residues 170 to 746 (referenced to strain H77) of the HCV polyprotein. Products were generated by nested polymerase chain reaction with genotype-specific primers (Table 1) using Expand high-fidelity polymerase (Roche). Each round of amplification used the thermal profile 94°C for 2 minutes, 25× (94°C for 45 seconds, 50°C for 45 seconds, 72°C for 90 seconds, extending 5 s/cycle) and 72°C for 7 minutes. Polymerase chain reaction products were cloned into the pCR3.1 vector (Invitrogen) according to the manufacturer's instructions. Clones were sequenced by using Big Dye chemistry (Applied Biosystems, Warrington, UK) and analyzed using an ABI PRISM 3100 sequencer. DNA sequences were aligned using ClustalX.26 Phylogenetic and molecular evolutionary analyses of gap-stripped sequences were estimated by the neighbor-joining method applied to pairwise distances estimated by the 2-parameter method. The reliability of the phylogenetic results was assessed using 1,000 bootstrap replicates. These analyses were performed using MEGA version 2.1.27

Table 1. Primers Used for Nested Amplification of HCV Glycoproteins, Amino Acids 170–746
Sense Primers  
Antisense primers  

Generation of HCVpp and Infection Assays.

HCVpp were produced as previously described13, 22 from 293T cells co-transfected with a murine leukemia virus (MLV) Gag-Pol packaging construct, an MLV-based transfer vector encoding a marker protein (GFP [green fluorescent protein] or beta-galactosidase), and the E1E2 expression constructs. Infection of target cells was performed as described previously.13, 22 As shown below, within a given preparation of virions, similar amounts of virion-associated MLV capsid proteins were detected for the pseudo-particles generated with the different HCV glycoproteins. However, important differences in the absolute quantities of virion-associated capsid proteins could be noticed when two independent preparations of pseudo-particles were compared (data not shown), despite comparable infectious titers (see below). Thus, to minimize artifacts caused by differences in the quality of preparations, each subsequent evaluation experiment was conducted using pseudo-particles generated concurrently. Moreover, because the detection of virion-associated capsid proteins did not appear to be a valid indicator of infectious particles and precluded comparison of results, normalization of the pseudotyped vector stocks was performed using the infectious titers determined on Huh-7 cells and expressed as TU (transducing units) per milliliter of producer cell supernatant.

RNA Interference Assay.

The three siRNAs directed against SR-BI28 mRNA were 5′-GCAGCAGGUCCUUAAGAAC (SR-BI/siRNA-2), 5′-GGACCCCCUUGUGAAUCUC (SR-BI/siRNA-23) and 5′-GGUU- GACUUCUGGCAUUCC (SR-BI/siRNA-34). They were expressed in target cells through a VSV-G–pseudotyped bicistronic lentiviral vector, FG12,29 allowing the expression of the siRNA and a GFP marker protein to control the transduction efficiencies.


Assembly and Host-Range of Infectious HCVpp Harboring E1E2 Glycoproteins of Different Genotypes.

To study the cell entry properties of the E1E2 glycoproteins from different HCV genotypes, we derived HCVpp-harboring E1E2 glycoproteins from subtypes 1a, 1b, 2a, 2b, and 3a, and genotypes 4, 5, and 6. A total of 88 HCV sequences were recovered from sera of patients infected with different genotypes, as determined by the genotyping INNO-LiPA HCV II method (Innogenetics, Ghent, Belgium). The sequences encoding the E1E2 envelope glycoproteins were then amplified from positions corresponding to amino acids 170 to 746 of the HCV polyprotein, cloned into a cytomegalovirus promoter-driven expression vector,13 and sequenced. Phylogenetic analysis, performed using reference strains, confirmed that genotype definition based on the sequence of the envelope genes concurred with the predicted genotype (Fig. 1). To generate infectious HCVpp, the E1E2 expression vectors were co-transfected into 293T cells with a Gag-Pol expression vector encoding MLV retroviral cores and with an MLV-derived transfer vector encoding the GFP marker protein. Although a limited number of the E1E2 sequences turned out to be functional in infection assays (i.e., 24 of the 88 clones), most expressed E2 protein in transfected cells, and, more importantly, we were able to derive infectious HCVpp for all of the major different genotypes and subtypes. Highly infectious titers, higher than 5 × 104 TU/mL, were readily obtained for HCVpp of genotypes/subtypes 1a, 1b, 2a, 2b, 4, 5, and 6 on human hepatoma Huh-7 cells (Table 2). The infectious titers of HCVpp of genotype 3 were lower by more than 1 log on the same target cells, yet we were unable to identify a more permissive cell type (Table 2). All liver cell lines tested, including Huh-7, PLC/PRF/5, and Hep3B, could be infected with HCVpp of the different genotypes, with the exception of HepG2 cells, which do not express CD81. Furthermore, none of the nonhepatic human cell lines tested, including osteosarcoma, epithelial adenocarcinoma, rhabdomyosarcoma, glioblastoma, and B- or T-lymphoid cell lines, were permissive to HCVpp (Table 2). Likewise, human T and B primary cells could not be infected (data not shown). To investigate the phenotypic properties conferred by diverse HCV glycoproteins, a subset of 10 infectious clones that were representative of each genotype and main subtype, as determined by their location in the phylogenetic tree (Fig. 1), were chosen for further study. A more limited characterization of the other clones was performed and confirmed that each of the 10 clones selected was representative of its clade (data not shown).

Figure 1.

Neighbor-joining tree of nucleotide sequences encoding all functional HCV E1 and E2 genes (amino acids 170-746, numbering based on AF011752) derived from sera of chronically infected individuals. Clone pre-fixes are designated according to the country and city of origin (UKN; Nottingham, United Kingdom), patient number, then clone number. All 23 clones expressed HCV E2 glycoproteins, as demonstrated using the AP33 monoclonal antibody.50 The following reference strains were included in the comparison: genotype 1a (AF011752), 1b (D01171; AF333324; D90208), 2a (D00944), 2b (AB030907), 3a (D17763), 3b (D49374), 4 (Y11604), 5 (Y13184), and 6 (Y12083). Sequences of predominant viruses retrieved from some patients have been submitted to GenBank under accession numbers AY734967 to AY734988, AY785283, and AY736194 for sequences UKN1A14.16, UKN1A14.43, UKN1A14.44, UKN1A14.18, UKN1A14.38, UKN1A14.42, UKN1A2.1, UKN1B12.16, UKN1B12.6, UKN1B5.23, UKN2A1.2, UKN2A1.5, UKN2A2.4, UKN2A2.5, UKN2A3.12, UKN2B1.1, UKN2B2.8, UKN3A.1.28, UKN3A1.9, UKN4.11.1, UKN4.21.16, UKN4.21.17, UKN5.14.4, and UKN6.5.340. Numbers along each branch indicate the number of bootstrap replicated trees that support that branch. The scale bar represents the rate of nucleotide substitutions per nucleotide site. E1E2 clones selected for in-depth phenotypic analysis are shown underlined and in bold type.

Table 2. Tropism of HCVpp Containing E1–E2 From Genotypes 1a, 1b, 2a, 2b, 3c, 4c, 5, and 6
Cell LinesSpeciesTypesGenotypes
H77CGUKN1B 12.6UKN2A 1.2UKN2A 2.4UKN2B 2.8UKN3A 1.28UKN4 11.1UKN5 14.4UKN6 5.340
  1. NOTE. The infectivity of HCVpp harboring the E1E2 glycoproteins of the indicated HCV genotypes and subtypes was detected on different cell types. (++), infectious titers higher than 5 × 104 TU/mL; (+), titers between 5 × 103 and 5 × 104 TU/mL; (+/−), titers between 5 × 102 and 5 × 103 TU/mL; (−), titers lower than 5 × 102 TU/mL, which corresponds to the threshold of detection of infected cells by FACS analysis, as determined using pseudo-particles generated in the absence of viral glycoproteins. The results were derived from at least three independent experiments using different virion production batches, and the standard deviations did not exceed 30% of the mean values. The infectivity of control pseudo-particles generated with the VSV-G envelope glycoprotein ranged from 5 × 106 to 2 × 107 TU/mL. Overexpression of SR-BI in 293 cells resulted in up to 10-fold increased infectious titers of HCVpp.

Huh-7humanhepatocellular carcinoma+++++++++++++++++++
Huh-7 L-SIGNhumanhepatocellular carcinoma+++++++++++++++++++
Hep3Bhumanhepatocellular carcinoma+++++++/−+++
HepG2 CD81humanhepatocellular carcinoma++++++++/−+++
HepG2humanhepatocellular carcinoma
SW-13humanadreno-cortical carcinoma
293 SR-BIhumanembryo kidney+/−+/−+/−+/−+/−+/−+/−+/−+/−
293humanembryo kidney
Helahumancervix adenocarcinoma
Hela L-SIGNhumancervix adenocarcinoma
Molt-4humanT-lymphoblastic leukemia
THP L-SIGNhumanB-lymphoma
THP DC-SIGNhumanB-lymphoma
A431humanepidermal carcinoma
COS-7monkeykidney fibroblasts
CHOhamsterChinese hamster ovary
CHO CD81/SR-BIhamsterChinese hamster ovary

Cell Entry Pathways.

On receptor binding, entry of enveloped viruses may occur either by direct membrane fusion at the cell surface, for the so-called pH-independent viruses, or after internalization in acid-pH endosomal vesicles, for pH-dependent viruses. To discriminate the cell entry pathway used by the different HCVpp genotypes, we treated Huh-7 target cells with different concentrations of bafilomycin A1, a specific inhibitor of the vacuolar proton pump adenosine triphosphatase involved in acidification of endosomes. As expected, infectivity of control pseudo-particles generated with the envelope glycoprotein of the RD114 retrovirus that uses a pH-independent membrane fusion mechanism was not affected by bafilomycin A1 (Fig. 2). In contrast, bafilomycin A1 inhibited the infectivity of HCVpp of all genotypes as well as of control pseudo-particles harboring HA or VSV-G glycoprotein that enter the cells through a pH-dependent pathway.30 Thus, these results suggest that HCV cell entry route proceeds by endocytosis and this process is conserved within the different HCV strains.

Figure 2.

pH-dependence of HCVpp cell entry. Results of infection assays performed using Huh-7 cells incubated with varying concentrations (nmol/L) of bafilomycin A1 (Sigma-Aldrich, Saint Quentin Fallavier, France) before (1 hour) and during infection (3 hours). Similar inputs of viral particles were used in the experiments, allowing infection of approximately 5% to 10% of target cells for the lower dilutions of viral supernatants. The results are expressed as percentages of inhibition (mean ± SD; n = 5) relative to titers determined in the absence of bafilomycin A1. FPV-HA, hemagglutinin of fowl plague virus; G, VSV-G, glycoprotein of vesicular stomatitis virus.

CD81 Is Critically Involved in the Entry Process of All HCV Genotypes.

A functional role for the CD81 tetraspanin and the SR-BI in cell entry was confirmed by using infectious HCVpp of genotype 1.13, 22, 23, 31, 32 We sought to investigate the conservation of usage of either receptor in cell entry of other HCV genotypes/subtypes. First, we expressed CD81 in HepG2 human hepatocarcinoma cells. In contrast to the parental cells, CD81-transfected HepG2 cells supported infection by HCVpp of all genotypes/subtypes, indicating that CD81 is involved in the cell entry process of all HCV genotypes (Fig. 3A). Second, to examine the requirement of CD81 at the early stages of HCV cell entry, we performed receptor-competition assays using anti-CD81 monoclonal antibodies that block binding of soluble E2 (Fig. 3C). The infectivity measured on Huh-7 cells was specifically inhibited by the antibodies, with inhibitions ranging between 70% and 99%. No difference in the inhibition of HCVpp from the different genotypes was detected, confirming the conserved usage of CD81 among HCV genotypes.

Figure 3.

Results of infection assays. (A) Complementation assays in HepG2 cells. The results of infection performed with Huh-7, HepG2, and HepG2-CD81 target cells are displayed as TU/mL supernatant (mean ± SD; n = 6). Control experiments were performed by using pseudotype particles generated with VSV-G glycoprotein, RD114 envelope glycoprotein, or lacking glycoproteins (no env). (B) Immunoblots of viral particles pelleted through 20% sucrose cushions are shown. The MLV capsid protein was revealed in reducing and denaturing conditions using an anti-capsid (MLV CA) antiserum (ViroMed Biosafety Laboratories, Minnetonka, MN), as described previously.13 (C) Results of CD81 receptor competition assays. Huh-7 cells were pre-incubated for 30 minutes with 30 μg/mL of JS-81, a CD81-blocking antibody (BD Pharmingen), before infection with HCVpp. Similar inputs of viral particles were used in the experiments, allowing infection of approximately 5% to 10% target cells for the lower dilutions of viral supernatants. Results are expressed as percentages of inhibition of the infectious titers obtained on Huh-7 cells relative to titers obtained in the absence of inhibitors (mean ± SD; n = 5). Efficient neutralization was also obtained with other CD81-blocking antibodies (data not shown). Control experiments were performed using pseudo-particles generated with VSV-G or RD114 glycoproteins.

Variable Implication of SR-BI in Cell Entry.

To investigate the conservation of SR-BI usage in HCVpp infection, we silenced its expression in Huh-7 cells by RNA interference. Three siRNAs that target different regions of the SR-BI mRNA were designed and expressed in target cells through a vesicular stomatitis virus (VSV)-G–pseudotyped human immunodeficiency virus (HIV)-1–based retroviral vector that also encoded the GFP marker protein. On infection with this vector, more than 99% of the Huh-7 cells expressed the GFP (data not shown), correlating well with SR-BI downregulation in most cells of the populations transduced with two of the three SR-BI siRNAs (Fig. 4 A). The reduction of SR-BI expression levels was of approximately 10-fold, as shown both by fluorescence-activated cell sorter analysis (Fig. 4A) and by Western blot analysis (Fig. 4B). We then evaluated the infectivity of HCVpp harboring E1E2 glycoproteins from various genotypes/subtypes on the SR-BI–downregulated Huh-7 cells (Fig. 4C). The SR-BI silencing did not inhibit the infectivity of control pseudo-particles generated with the RD114 envelope glycoprotein, as expected, but induced variable effects on infectivity of HCVpp of different genotypes, even if all were sensitive to SR-BI down-regulation (Fig. 4C). The HCVpp harboring E1E2 glycoproteins derived from strain H77 were the most sensitive to SR-BI down-regulation (titers reduced by more than 90%). Conversely, the infectivity of HCVpp derived from a 2a subtype was three-fold to four-fold less efficiently inhibited by SR-BI silencing than HCVpp of 1a genotype. Likewise, all genotypes could be neutralized by an SR-BI blocking serum22 in a dose-dependent fashion, and the same pattern of sensitivity to SR-BI blocking could be demonstrated (Fig. 4D).

Figure 4.

Inhibition of SR-BI–mediated HCV-pp infection. Downregulation of SR-BI in Huh7 cells was achieved using a VSV-G–pseudotyped lentiviral vector, FG12,29 encoding siRNAs directed against SR-BI and the GFP marker protein. (A) Permeabilized Huh-7 cells were stained with an SR-BI rabbit anti-serum (ab396, Abcam) (solid lines) or with a rabbit control serum (dashed lines) and analyzed by flow cytometry. (B) Expression of SR-BI in Huh-7 cells transduced with siRNAs was analyzed by Western blot analysis using a SR-BI rabbit anti-serum (ab396, Abcam; 1:1,500). (C) Results of infection in SR-BI–downregulated Huh-7 cells with the indicated pseudo-particles harboring a lacZ marker gene are expressed as percentages of inhibition of the infectious titers relative to titers obtained in parental Huh-7 cells (mean ± SD; n = 4). Infection of Huh-7 cells transduced with the FG12 empty vector raised the same results as parental Huh-7 cells. (D) Results of infection of Huh-7 target cells with the indicated pseudo-particles performed in the presence of a SR-B1 blocking serum at a 1:50 dilution. The results are expressed as percentages of inhibition (mean ± SD; n = 3) relative to titers determined in the presence of control serum. Control experiments were performed using pseudo-particles generated with the RD114 glycoprotein. nd, not determined.

Cross-Neutralization of HCVpp Cell Entry by Human Sera.

Antibodies directed against the surface glycoproteins of enveloped viruses can neutralize infection by inhibiting receptor binding and post-binding steps. Such antibodies can be raised during infection but may not inhibit cell entry mediated by variant glycoproteins. Thus, to evaluate the extent of cross-neutralization, we performed neutralization assays using HCVpp and sera from chronic HCV carriers (Fig. 5). No neutralization of control pseudo-particles could be detected with these patient sera. A first serum from a patient infected with a subtype 1b HCV (S1) could neutralize the infectivity of the HCVpp harboring the E1E2 glycoproteins of the different genotypes/subtypes; yet with different efficiencies. However, an alternative genotype 1b serum (S7) did not generate as broad neutralization and raised strain-specific neutralization responses in some cases (e.g., gt3a and gt4). Likewise, a patient serum harboring a genotype 3 (S11) displayed both genotype- and strain-restricted neutralization responses. Sera from other chronic carriers with defined HCV genotypes/subtypes raised similar results (data not shown); that is, broadly reactive neutralization activities that could not always inhibit cell entry of the full range of genotypes and subtypes tested.

Figure 5.

Neutralization of HCV pseudo-particles. Results of neutralization of HCVpp pre-incubated before infection of Huh-7 cells with decomplemented sera from chronically HCV-infected patients. Negative control experiments were performed using no antibodies or using pseudo-particles generated with VSV-G or RD114 glycoprotein. Similar inputs of viral particles were used in the experiments, allowing infection of approximately 5% to 10% target cells for the lower dilutions of viral supernatants. Results are expressed as the percentages of inhibition of the infectious titers (mean ± SD; n = 4) relative to incubation in the presence of a pool of naive human sera (n = 4). HCVpp were pre-incubated for 30 minutes at room temperature with sera diluted 1:50 before infection of Huh-7 target cells. The predominant genotype of HCV diagnosed in the patients is indicated in parentheses. Efficient neutralization of the control pseudo-particles (data not shown) was demonstrated using a hyper-immune goat serum raised against the RD114 SU glycoprotein (ViroMed Biosafety Laboratories).


Generation of Infectious HCVpp Harboring E1E2 Glycoproteins of Different Genotypes/Subtypes.

In this study, we investigated the cell entry properties from different E1E2 clones representative of the diverse HCV genotypes and subtypes. Although approximately 25% of E1E2 sequences appeared to be functional, we were able to isolate functional envelope glycoproteins from genotypes/subtypes 1a, 1b, 2a, 2b, 3a, 4, 5, and 6 that confer infectivity to HCVpp harboring a marker gene. We show that the HCVpp from the different genotypes/subtypes displayed a similar tropism for the human hepatoma cells. Interestingly, differences in the infectious titers of the HCVpp were detected. Whereas HCVpp generated with E1E2 glycoproteins of gt1a, gt1b, gt2a, or gt2b had high titers, in the range of 105 TU/mL, the infectious titers of HCVpp harboring glycoproteins of gt3a were generally lower, by more than 1 log. Whether these differences of infectivity were attributable to an inadequate choice of the target cells used in the infection assays, to sequence artifacts incorporated during the reverse transcription polymerase chain reaction (RT-PCR) steps, to retrieval of non-competent variants present in the quasispecies from patients' samples, to poor viral incorporation of some of the glycoproteins on HCVpp, or to other intrinsic features of these HCV glycoproteins remains to be determined. The recovery of a wider range of isolates will establish whether these differences in infectivity are real. However, in the lines of previous results obtained with E1E2 glycoproteins from prototype 1a and 1b HCV isolates,13, 22, 23 our results indicate that hepatocytes may be the preferred HCV target cells in vitro because the E1E2 glycoproteins from all of the different HCV genotypes tested efficiently mediated cell entry only into hepatocarcinoma cells. We were unable to detect HCVpp entry into nonhepatic cell lines from several different tissues (including B- or T-lymphoid cell lines) as well as in resting or stimulated peripheral blood mononuclear cells. Therefore, these results are not fully consistent with in vivo data that suggest that HCV may infect a wider range of cell types, including B cells, monocytes/macrophages, and dendritic cells,33–35 as well as nonhematopoietic cell types.36 The HCV clones used in this study were retrieved from sera of HCV-infected patients; this may induce a bias for hepatocytes. Alternatively, the HCV variants that infect extrahepatic targets may poorly replicate in patients and thus may not be easily cloned. The characterization of additional HCV sequences from viruses isolated from alternative cell types, including peripheral blood mononuclear cells and dendritic cells, will help to determine the range of HCV tropism in vivo.

Characterization of a Common Entry Pathway Between Genotypes/Subtypes.

We demonstrate that HCVpp-harboring envelope glycoproteins from the different genotypes/subtypes enter the cells by the endocytic pathway and require the acidification of endosomal vesicles to trigger the fusogenic activity of their fusion proteins. This result is expected owing to the structural homology between the hepacivirus glycoproteins and their classification in Flaviviridae, a family of viruses whose cell entry routes are pH dependent for all members. Moreover, by receptor competition assays with CD81-blocking antibodies as well as by ectopic expression of CD81 in nonpermissive hepatic cells, we found that the CD81 tetraspanin is used by all genotypes and subtypes tested to promote cell entry. Such a conserved usage of CD81 as the cell entry receptor is surprising, because several binding studies using soluble forms of the E2 protein (sE2) indicated considerable discrepancies between genotypes in terms of affinity to bind CD8120; sE2 from some genotype 3 was unable to bind CD81.25 Thus, these results show that the strength of the interaction between sE2 and CD81 does not predict the CD81 dependence of infection by HCV genotypes, as also recently pointed out in infection assays using CD81 binding mutants with reduced affinity to sE2.32 The functional interaction of HCV with CD81 and other cellular molecules seems more complex than the interaction of monomeric sE2 with CD81.

Variations in SR-BI Dependence for Infectivity.

The SR-BI dependence for HCVpp infection was evaluated by RNA interference, an approach that has proved powerful to study the function of viral receptors during HIV infection.29 The downregulation of SR-BI reduced the infection of all HCVpp tested, consistent with the augmented infectivity of all HCVpp in 293 cells overexpressing SR-BI (Table 2). However, we could discriminate the different isolates by the levels of inhibition induced by SR-BI silencing or blocking (Fig. 4). Indeed, whereas the infectivity of HCVpp of subtype 1a was reduced by more than 10-fold, the infectivity of one of the 2a subtype (UKN2A 1.2) was marginally reduced, by less than 30%. Despite this difference, it is clear that CD81 and SR-BI are both required for infection but are not sufficient to mediate HCVpp infection in vitro, because SR-BI/CD81-transfected CHO cells were unable to support infection by all HCVpp genotypes and subtypes (Table 2). Our results also indicate that an unknown hepatocyte-specific component is involved in a receptor complex that encompasses CD81 and SR-BI, because several human cells (e.g., HeLa and HOS cells22) co-expressing both receptors are nonpermissive (Table 2). The varying dependence of HCVpp genotypes/subtypes to SR-BI suggests that the stoichiometry of the molecules involved in an optimal receptor complex may differ among HCV genotypes, subtypes, or isolates. Alternatively, these results may indicate that the levels of SR-BI expression in siRNA-treated Huh-7 cells are not sufficient to promote efficient entry of some HCV genotypes. Indeed, variations in SR-BI receptor usage may be inherent to a differential capacity of SR-BI to interact with the E1E2 glycoproteins for some genotypes, so that SR-BI would mediate efficient infection only when expressed at high density. Several reports37, 38 have established that potentially active receptors fail to mediate γ-retrovirus infections when they are expressed under a certain threshold of expression. Similarly, a mutant HIV-1 co-receptor, CCR5, with a large reduction in affinity for HIV-1 glycoprotein, was found inactive at low concentrations, but was almost as active as wild-type CCR5 at high concentrations,39 implying the involvement of multiple co-receptors in the HIV-1 receptor complex. Further studies using cell lines expressing different amounts of SR-BI or CD81 receptors will be necessary to determine the number of molecules that are needed to trigger HCV entry. Moreover, identifying the modular element of E1E2 complex that is capable of adapting to changing concentrations of SR-BI will be important. Alternatively, we cannot exclude that the variation in SR-BI dependence for infectivity is attributable to alternative receptor molecules that might be differentially used by different HCV genotypes/subtypes or clones. Several other molecules, for example, DC-SIGN, L-SIGN,40, 41 and the asialoglycoprotein receptor,42 have been shown to bind to HCV E1E2 glycoproteins or to pseudo-particles, but it seems clear that their role is limited to the capture of viral particles.22, 23, 41 Efforts to identify other liver-specific molecules will provide insights in the initial steps of HCV infection. The understanding of the multi-step process of infection should offer good opportunities for therapeutic intervention.

Cross-Neutralization of HCVpp Cell Entry.

Several epidemiological and experimental studies provide evidence that polyclonal antibody to HCV can protect from hepatitis C. HCV immunoglobulins or hyperimmune sera can indeed delay or even prevent hepatitis C in human and chimpanzees, when the virus is inoculated after or co-inoculated with the antibodies.43–45 Such reagents contain high-titer antibodies against E1E2 glycoproteins and can inhibit the infectivity of HCVpp in vitro in a complement-independent manner,13, 45, 46 hence suggesting neutralization mechanisms at the level of cell entry. Additionally, chimpanzees vaccinated with recombinant HCV glycoproteins that induced high-titer antibodies were partially protected against a subsequent low-dose homologous HCV challenge.47 However, the demonstration that both experimentally infected chimpanzees and naturally infected humans could be reinfected with heterologous HCV strains48 and that chronic infection correlated with increased viral genetic evolution, particularly in the HVR1 region,49 suggested a narrow cross-neutralization capacity. Our results obtained with a limited number of sera from HCV-genotyped chronic carriers confirm the presence of antibodies directed against E1E2 glycoproteins that can neutralize cell entry, most likely by inhibiting receptor binding and post-binding steps, and indicate that whereas some sera display broadly reactive neutralizing activities, others seem to react in both genotype and strain-restricted manners. Further characterization of the neutralization of infection is crucial to understand HCV pathogenesis and to develop vaccines. Thus, the HCVpp described in this report will provide a powerful system to investigate the mechanisms of HCV neutralization and cross-neutralization.


The authors thank Pascale Nony and Sophie Chappuis from TRANSAT for their help in producing siRNA lentiviral vectors; Pierre-Yves Lozach and Ralf Altmeyer for providing L-/DC-SIGN vectors; Christelle Granier for excellent technical assistance; and all of the laboratory members for stimulating discussions.