Hepatitis C virus comes full circle: Production of recombinant infectious virus in tissue culture

Authors

  • Jan Martin Berke,

    1. Division of Gastroenterology and Hepatology, Centre Hospitalier Universitaire Vaudois, University of Lausanne, Lausanne, Switzerland
    2. Department of Medicine II, University of Freiburg, Freiburg, Germany
    Search for more papers by this author
  • Darius Moradpour

    Corresponding author
    1. Division of Gastroenterology and Hepatology, Centre Hospitalier Universitaire Vaudois, University of Lausanne, Lausanne, Switzerland
    2. Department of Medicine II, University of Freiburg, Freiburg, Germany
    • Division of Gastroenterology and Hepatology, Centre Hospitalier Universitaire Vaudois, Rue du Bugnon 46, CH-1011 Lausanne, Switzerland
    Search for more papers by this author
    • fax: (41) 21-314-47-18


  • Potential conflict of interest: Nothing to report.

Hepatitis C virus (HCV) has frustrated researchers since its identification in 1989.1 The virus could not be conclusively visualized, biochemical characterization of native viral products was precluded by the low titers in serum and liver tissue, and—most importantly—HCV could not be efficiently cultured in vitro, impeding the study of the viral life cycle as well as the development of specific antiviral agents and preventive vaccines. Despite these obstacles, formidable progress has been made over the last 16 years using heterologous expression systems, functional complementary DNA clones that are infectious in vivo in chimpanzees,2 the replicon system,3 and, more recently, functional HCV pseudoparticles that allow examination of viral entry under reproducible and conveniently measurable conditions.4, 5

A milestone has now been reached, with three groups reporting the production of recombinant infectious HCV in tissue culture.6–8 The purpose of this concise review is to highlight this important breakthrough. The interested reader is referred to more comprehensive review articles for a detailed discussion of other recent advances in the investigation of the molecular virology and pathogenesis of hepatitis C.9–15

Studying the Viral Particle and Its Entry Pathway

The basic biophysical properties of the HCV particle were revealed early on through experiments performed in chimpanzees. Infectivity was abolished by treatment with lipid solvents, indicating that the viral particle is enveloped.16 In addition, an estimate of the virion size was obtained via filtration studies demonstrating that the particle has a diameter of approximately 50 nm.17 Subsequent electron microscopy studies confirmed that HCV is an enveloped spherical particle that is 40-70 nm in diameter.18, 19 Thus, it is believed that the HCV envelope contains the viral glycoproteins E1 and E2 and encloses a nucleocapsid composed of multiple copies of core protein, which in turn harbors the viral positive-strand RNA genome.

Candidate HCV receptors have been identified, including CD81, a tetraspanin molecule found on the surface of many cell types,20 the low-density lipoprotein receptor,21 and scavenger receptor class B type I.22 However, validation of these and other candidates has been limited by the paucity of experimental systems for analyses of the early steps of the viral life cycle.

Infection of established cell lines and primary hepatocytes with serum-derived HCV in vitro yielded only low-level replication, requiring quantitative reverse-transcriptase polymerase chain reaction techniques as an assessment of successful infection.23, 24 Various alternatives have therefore been explored. Soluble C-terminally truncated versions of E2,20, 22 liposomes reconstituted with E1 and E225 and virus-like particles expressed in insect cells26, 27 have been used to study HCV glycoprotein interactions with the cell surface. More recently, the production of virus-like particles28 and putative virions29 has also been described in mammalian cells. In addition, pseudotyped vesicular stomatitis virus or influenza virus particles have been produced, incorporating E1 and E2 glycoproteins, whose C-terminal transmembrane domains were modified to allow transport to the cell surface.30 However, such modifications may interfere with the complex roles of the E1 and E2 transmembrane domains and may perturb the conformation and functions of E1–E2 complexes.

On this background, the establishment of infectious retroviral pseudotypes displaying functional HCV glycoproteins as a robust model system for the study of viral entry represented an important breakthrough.4, 5 HCV pseudoparticle infectivity is restricted primarily to human hepatocytes and hepatocyte-derived cell lines, can be blocked by antibodies against E2 or sera from patients with hepatitis C, and is pH-dependent. Thus, HCV entry likely involves transit through an endosomal, low-pH compartment and fusion with the endosomal membrane, a process that has recently been elucidated in great detail in related flaviviruses.31, 32 Retroviral pseudotypes allowed, for example, to validate CD81 as an essential component of the HCV entry pathway33–35 and to investigate the role of neutralizing antibodies in acute and chronic hepatitis C.36–39 Despite these advances, robust recapitulation of the entire HCV life cycle in tissue culture remained elusive.

Abbreviation

HCV, hepatitis C virus.

The Replicon System

The development of a replicon system for HCV by Ralf Bartenschlager and colleagues now at the University of Heidelberg, Germany, revolutionized the investigation of HCV RNA replication3 (see the review by Bartenschlager40). The prototype subgenomic replicon was a bicistronic RNA in which the structural region and part of the nonstructural region of HCV were replaced by the neomycin phosphotransferase gene and translation of the nonstructural proteins 3-5B was driven by a second, heterologous internal ribosome entry site from encephalomyocarditis virus. Using this approach, it became possible to study efficient and genuine HCV RNA replication in Huh-7 human hepatocellular carcinoma cells in vitro. Interestingly, certain amino acid substitutions (i.e., cell culture–adaptive mutations) were found to increase the efficiency of replication by several orders of magnitude.41, 42 Such adaptive mutations cluster in certain regions, such as the central region of nonstructural protein 5A (NS5A), the C-terminal portion of the NS3 serine protease and the N-terminal portion of the NS3 RNA helicase domains as well as two positions in NS4B.43 In addition, the efficiency of replicon RNA amplification was found to be determined by selection for particularly permissive cells within a given population of Huh-7 cells.43, 44 Accordingly, removal of a replicon from a given cell clone by treatment with interferon α or a selective drug often results in cell lines that support higher levels of HCV RNA replication as compared with naïve Huh-7 cells. In one prominent example, an interferon α–“cured,” Huh-7–derived cell line designated Huh-7.5,44 a single point mutation in the double-strand RNA sensor retinoic acid–inducible gene I (RIG-I) appears to determine higher permissiveness for HCV RNA replication.45 In most other cases, however, the molecular basis for increased permissiveness is unknown.

The replicon system has allowed genetic dissection of HCV RNA elements and proteins, provided material for biochemical and ultrastructural characterization of the viral replication complex, and facilitated drug discovery efforts as well as the investigation of antiviral resistance. Moreover, the replicon system has been exploited for analyses of the effect of cytokines on HCV RNA replication46, 47 and was instrumental in a recent series of elegant studies looking at host proteins involved in HCV RNA replication as well as mechanisms of evasion from innate immune responses.48–53

Since the original reports of functional genotype 1b replicons, replicons for genotype 1a54 and 2a55 as well as derivatives expressing easily quantifiable marker enzymes in a separate cistron have been made to facilitate genetic studies as well as drug screening and evaluation.56–58 In addition, full-length replicons and HCV genomes efficiently replicating in tissue culture have been developed,44, 59, 60 and the spectrum of permissive host cells has been expanded.61 Finally, replicons have been established that allow tracking of functional HCV replication complexes in living cells.62

The JFH-1 Clone As a Turning Point

One puzzling and disappointing observation was that full-length HCV genomes with adaptive mutations were incapable of producing infectious virus. It was thought that either the host cell (i.e., the Huh-7 cell line) lacked some factors critical for particle formation and release or that the adaptive mutations required for efficient replication in tissue culture interfered with packaging, assembly or release of virus. In support of the second idea, an inverse correlation was found between mutations that permit efficient replication of HCV RNA in HuH-7 cells in vitro and productive replication after intrahepatic inoculation into chimpanzees in vivo.63 Thus, one possibility is that cell culture adaptive mutations that promote efficient RNA replication in vitro may be inefficient in vivo because they compromise particle assembly and release.

An unexpected twist came when Takaji Wakita and his colleagues from the Tokyo Metropolitan Institute of Neuroscience reported an HCV genotype 2a clone isolated from a patient with fulminant hepatitis C, designated as JFH-1 (Japanese fulminant hepatitis 1) that replicated in Huh-7 cells as well as in other liver-derived64 and nonhepatic cell lines65 without the requirement for adaptive mutations. This observation laid the foundation for a long-awaited breakthrough in HCV research. Indeed, three groups have now reported that cloned JFH-1 genomes transfected into Huh-7 cells not only replicate efficiently but produce virus that is infectious for naive Huh-7 cells, allowing study of the complete life cycle of HCV in vitro.

Production of Recombinant HCV in Tissue Culture

Takaji Wakita and Ralf Bartenschlager set the stage at last year's 11th International Symposium on HCV and Related Viruses in Heidelberg when their groups reported that either full-length JFH-1 RNA or chimeras comprising the core to NS2 region from the genotype 1b clone Con1 and the nonstructural region from JFH-1 produced infectious virus in Huh-7 cells. Their data were published in part in the July 2005 issue of Nature Medicine.7 Charles M. Rice and colleagues at The Rockefeller University in New York succeeded in establishing a highly efficient replication and virus production system by constructing chimeras comprising the core to NS2 region from the genotype 2a clone J6 and the nonstructural region from the JFH-1 clone. Their results were published in the July 22, 2005 issue of Science.6 Francis V. Chisari and colleagues at the Scripps Research Institute in La Jolla took advantage of the full-length JFH-1 clone and of a derivative of the highly permissive Huh-7.5 cell line to generate a robust virus production system, as reported in the June 28, 2005 issue of the Proceedings of the National Academy of Sciences.8

What are the results of these studies? Wakita et al.7 transfected in vitro–transcribed full-length JFH-1 RNA into Huh-7 cells and found that this genome not only replicated efficiently but produced virus that was infectious for naïve Huh-7 cells. Remarkably, infection could be tracked easily via immunofluorescence microscopy for viral proteins in infected Huh-7 cells. In this system, infection was found to spread in a limited fashion, and serial passages led to decreasing viral titers. Via immunoelectron microscopy using an E2-specific monoclonal antibody, the authors succeeded in demonstrating the presence of specifically labeled spherical particles measuring 50-65 nm in diameter in the culture supernatant from JFH-1 RNA-transfected cells, confirming the early filtration and chimpanzee transmission as well as electron microscopy studies mentioned above.17–19 Furthermore, Wakita et al. showed that viral particles had a density of approximately 1.17 g/mL and that infection could be blocked by a monoclonal antibody directed against CD81, confirming the importance of this molecule as a component of the HCV entry pathway. The authors also developed a bicistronic luciferase reporter construct with similar capacity to produce infectious particles as a conveniently measurable read-out. Using this elegant tool, they found that sera from patients with hepatitis C contained neutralizing antibodies that could block infection by recombinant HCV. Finally, as a proof-of-principle, they demonstrated that HCV particles produced in vitro are infectious when intravenously inoculated into a chimpanzee.

Lindenbach et al.6 generated chimeras comprising the core to NS2 region of either the infectious genotype 2a strain J6 or the genotype 1a strain H77 and the nonstructural region of JFH-1. Transfection of the J6/JFH-1 but not the H77/JFH-1 chimera into Huh-7.5 cells resulted in the efficient release of viral particles with peak infectivity titers of 105 tissue culture infectious units per milliliter 48 hours after transfection. Gradient analyses revealed a broad peak of infectivity, with the most infectious material found at a density of 1.09-1.10 g/mL, which is similar to the peak of infectivity seen in HCV-infected chimpanzees.66 In this system, infection spread efficiently from cell to cell and could be serially passaged without loss of infectivity titers. Interestingly, the ability of the H77/JFH-1 genotype 1a/2a chimera to replicate but not to spread suggests that interactions between the structural and nonstructural gene products may be important for infectious particle formation, as has previously been observed for other members of the Flaviviridae family.67, 68 Infection could be neutralized by an E2-specific monoclonal antibody and a soluble recombinant form of the CD81 large extracellular loop. In addition, HepG2 cells (another hepatocellular carcinoma cell line) engineered to express CD81 but not naïve HepG2 cells could be infected by the viral particles, confirming that CD81 plays an essential role in the HCV entry process. Finally, the virus produced in tissue culture could be inhibited efficiently by interferon α as well as specific HCV protease and polymerase inhibitors, demonstrating the suitability of this system for antiviral testing.

Zhong et al.8 derived an interferon γ–“cured” subclone, designated Huh-7.5.1, from the highly permissive Huh-7.5 cell line and demonstrated that infectious particles were released with high efficiency following transfection of full-length JFH-1 RNA. Infectivity titers were in the range of 104 to 105 infectious units per milliliter, which was approximately 50-fold higher as compared with the titers observed by Wakita et al.7 using parental Huh-7 cells. Virus spread efficiently throughout the culture and could be serially passaged without loss of infectivity. It was reasoned that the defect in the RIG-I pathway in Huh-7.5 cells and, by consequence, an impaired innate immune response against double-strand RNA generated during HCV RNA replication may be the reason for the higher efficiency. However, additional as-yet unidentified factors may contribute to the high permissivity of Huh-7.5 cells to HCV infection and replication. As pointed out in an accompanying editorial,69 careful analysis of the slow initial kinetics by which infectious virus is released from JFH-1 RNA-transfected Huh-7.5.1 cells raises the possibility that cell culture adapted JFH-1 variants emerged in this study. In agreement with this assumption, infection of Huh-7.5.1 cells with virus harvested from cells 10-24 days after transfection resulted in high-titer release of infectivity with much accelerated kinetics, suggesting that adapted variants were present in the inoculum. Clarification of this issue is likely to yield new insights into the complex interaction of HCV with the host cell. In addition, it will be interesting to investigate whether the efficient viral spread observed by Lindenbach et al.6 and Zhong et al.8 occurs via preferential attachment and entry into neighboring cells after secretion or via direct transmission from cell to cell.

The Next Frontiers

Taken together, these three papers represent a milestone in the field and remove a roadblock that has impeded HCV research for the last 16 years. Each step of the viral life cycle can now be studied, as is schematically illustrated in Fig. 1. Indeed, HCV begins to behave like a real virus, and classical virological techniques can now be applied to this important human pathogen, allowing us to gain insight into as-yet obscure steps of its life cycle, including viral entry, genome packaging, virion assembly, maturation, and release. Reverse genetics, in which distinct mutations introduced into the viral genome are analyzed for their impact on replication and virus production, will shed light onto the functions of hitherto difficult-to-study HCV proteins (e.g., the p7 polypeptide and the NS2 protein), and their possible role in particle formation can now be addressed. In addition, findings with implications for the pathogenesis of hepatitis C can be validated in a true infection system.70 Finally, novel antiviral strategies aimed at inhibiting the early or late steps of the viral life cycle will be explored using a robust system.

Figure 1.

Production of recombinant infectious HCV in tissue culture. Transfection of Huh-7 cells with in vitro–transcribed JFH-1 RNA leads to the production of viral particles that are infectious for naïve Huh-7 cells and can be serially passaged, thereby allowing study of the entire life cycle of HCV in tissue culture. Steps of the viral life cycle include (1) virus binding and internalization, (2) cytoplasmic release and uncoating, (3) internal ribosome entry site–mediated translation and polyprotein processing, (4) RNA replication in a membrane-associated replication complex, (5) packaging and assembly, and (6) virion maturation and release. The topology of HCV structural and nonstructural proteins at the endoplasmic reticulum membrane is shown schematically. HCV RNA replication occurs in a specific membrane alteration: the membranous web. Note that internal ribosome entry site–mediated translation and polyprotein processing, as well as membranous web formation and RNA replication—illustrated here as separate steps for simplicity—may occur in a tightly coupled fashion. ER, endoplasmic reticulum; MW, membranous web.

What are the next frontiers? One of the most important unresolved questions is: What renders the JFH-1 genome so special, and what elements in the nonstructural region affect viral particle production and release? Studies using chimeras and analyses of the behavior of the JFH-1 clone in vivo when inoculated into chimpanzees are likely to provide answers. In addition, such studies may clarify how faithfully the HCV particles produced in vitro from cloned genomes reflect authentic virions. If the determinants for efficient replication and virion production by the JFH-1 isolate can be mapped, it may become possible to extend this cell culture system to other virus isolates—notably of genotype 1, which is the most frequent worldwide. Further improvements may provide sufficient material for purification and biochemical analyses of viral particles and may ultimately yield high-resolution structures of the HCV virion, as has been obtained for other members of the Flaviviridae family.71, 72 Finally, robust experimental tools and an improved understanding of the factors responsible for HCV cell tropism and entry may facilitate the development of novel small animal models of HCV infection and pathogenesis. Clearly, exploitation of the novel HCV cell culture systems will yield insights into not only the life cycle of HCV but also fundamental host cell processes that may become amenable to antiviral intervention in the future.

Ancillary