Strain-dependent viral dynamics and virus-cell interactions in a novel in vitro system supporting the life cycle of blood-borne hepatitis C virus

Authors

  • Hussein Hassan Aly,

    1. Laboratory of Human Tumor Viruses, Institute for Virus Research, Kyoto University, Kyoto, Japan
    2. Hepatology Department, National Hepatology and Tropical Medicine Research Institute, Cairo, Egypt
    Search for more papers by this author
  • Yue Qi,

    1. Laboratory of Viral Oncology, Graduate School of Biostudies, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Kimie Atsuzawa,

    1. Department of Anatomy, Fujita Health University School of Medicine, Toyoake, Japan
    Search for more papers by this author
  • Nobuteru Usuda,

    1. Department of Anatomy, Fujita Health University School of Medicine, Toyoake, Japan
    Search for more papers by this author
  • Yasutsugu Takada,

    1. Department of Surgery, Division of Hepato-Pancreato-Biliary and Transplant Surgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Masashi Mizokami,

    1. Research Center for Hepatitis and Immunology, International Medical Center of Japan Kounodai Hospital, Ichikawa, Japan
    Search for more papers by this author
  • Kunitada Shimotohno,

    Corresponding author
    1. Center for Human Metabolomic Systems Biology, Keio University, Tokyo, Japan
    • Center for Human Metabolomic Systems Biology, Keio University, 35, Shinano-machi, Shinjuku-ku, Tokyo, 160-8582, Japan
    Search for more papers by this author
    • fax: 81-3-5363-3592

  • Makoto Hijikata

    Corresponding author
    1. Laboratory of Human Tumor Viruses, Institute for Virus Research, Kyoto University, Kyoto, Japan
    2. Laboratory of Viral Oncology, Graduate School of Biostudies, Kyoto University, Kyoto, Japan
    • Laboratory of Human Tumor Viruses, Institute for Virus Research, Kyoto University, 53, Kawaharacho, Shogoin, Sakyoku, Kyoto, 606-8507, Japan
    Search for more papers by this author
    • fax: 81-75-751-3998


  • Potential conflict of interest: Nothing to report.

Abstract

We developed an in vitro system that can be used for the study of the life cycle of a wide variety of blood-borne hepatitis C viruses (HCV) from various patients using a three-dimensional hollow fiber culture system and an immortalized primary human hepatocyte (HuS-E/2) cell line. Unlike the conventional two-dimensional culture, this system not only enhanced the infectivity of blood-borne HCV but also supported its long-term proliferation and the production of infectious virus particles. Both sucrose gradient fractionation and electron microscopy examination showed that the produced virus-like particles are within a similar fraction and size range to those previously reported. Infection with different HCV strains showed strain-dependent different patterns of HCV proliferation and particle production. Fluctuation of virus proliferation and particle production was found during prolonged culture and was found to be associated with change in the major replicating virus strain. Induction of cellular apoptosis was only found when strains of HCV-2a genotype were used for infection. Interferon-alpha stimulation also varied among different strains of HCV-1b genotypes tested in this study. Conclusion: These results suggest that this in vitro infection system can reproduce strain-dependent events reflecting viral dynamics and virus-cell interactions at the early phase of blood-borne HCV infection, and that this system can allow the development of new anti-HCV strategies specific to various HCV strains. (HEPATOLOGY 2009.)

Hepatitis C virus (HCV) is a serious problem worldwide, with 3% of the world's population chronically infected.1 Chronic infection with HCV may lead to high rates of liver cirrhosis and hepatocellular carcinoma.2 Because the HCV standard therapy is still insufficient for treating many patients,3 the development of more effective and less toxic anti-HCV agents is desired. The virological studies required to reach this goal need reproducible and efficient HCV proliferation in cell culture. An in vitro infection system using recombinant HCV-JFH1 was developed. In this system, HuH7 cells transfected with in vitro-synthesized JFH1-RNA were shown to secrete infectious viral particles.4 This system, however, requires the combination of HuH-7-derived cell lines and JFH1-based constructs, limiting its usefulness for studying other HCV strains. Because HuH-7 cells cannot support the complete life cycle of blood-borne HCV (bbHCV) derived from clinical samples,5 this system is insufficient for studying all the events related to bbHCV infection.

Many researchers have attempted to develop an in vitro system for bbHCV.6–8 These current systems, however, are still insufficient due to their low efficiency for infectivity and replication of bbHCV. Working toward this same goal, we recently established immortalized primary human hepatocyte cell lines by transducing them with E6 and E7 genes from the human papilloma virus 18.5, 9 As expected, we observed improved infection and replication of bbHCV especially in one of these cell lines (HuS-E/2 cells) that showed a similar expression profile to that of human primary hepatocytes, but this strategy did not improve production of infectious particles.

Recently, a hybrid artificial liver support system was developed using animal hepatocytes cultured in a three-dimensional hollow fiber (3D/HF) system. This bioartificial liver showed several characteristic features of liver tissue for more than 4 months.10–12 By growing our HuS-E/2 cells in a similar 3D culture5 the gene expression profile was improved to more closely match that of human primary hepatocytes. Because the 3D cell culture condition more closely reproduces the in vivo environment of hepatocytes,13 culturing these cells in this manner may support the entire HCV life cycle.

In this study we utilized this small 3D culture system and showed it to be ideal for culturing HuS-E/2 cells for the study of bbHCV infection. Using this system we are now able to study the variable patterns of the life cycle of different bbHCV strains as well as HCV-related cellular events.

Abbreviations

2D, two-dimensional; 2D-HuS-E/2, HuS-E/2 cells cultured in two-dimensional condition; 3D, three-dimensional; 3D/HF, 3D hollow fibers; 3D-HuS-E/2, HuS-E/2 cells cultured in three-dimensional condition in the hollow fibers; HCV, hepatitis C virus; IFN-α, interferon alpha; LDH, lactate dehydrogenase; p.i., postinfection; RFB, radial-flow bioreactor; RT-PCR, reverse transcription polymerase chain reaction.

Materials and Methods

Cell Culture.

HuS-E/2 cells were cultured as previously described.5 For the 3D/HF system, HuS-E/2 suspension was injected into the lumen of HF (Toyobo, Osaka, Japan) made from cellulose acetate and containing pores for nutrients and waste exchange (Supporting Fig. 1). The bundles were centrifuged to induce organoid formation. The cells in the fibers were cultured in 12-well plates (two capillary bundles per well) with gentle rotation using serum-free medium (Toyobo) in a CO2 incubator at 37°C. The number of cells was adjusted to 3 × 105 cells per two-capillary bundle at the start of each experiment.

RNA Experiments.

Total RNA was extracted from two-dimensional (2D) cultured cells, patient sera, or from 100 times concentrated culture medium as previously described.4, 5 For cells cultured in the 3D/HF, sterile scissors were used to cut each fiber into small pieces (1 mm2 each), which were then solubilized in Sepasol RNA-1 (Nacalai Tesque, Kyoto, Japan). RNA was then extracted according to the manufacturer's protocol. Real-time reverse transcription polymerase chain reaction (RT-PCR) was performed as described.5

HCV Infection.

HCV infection experiments were carried out using sera from HCV patients. The amount of each inoculum was adjusted so as to add similar amount of HCV-RNA to the medium of the cells. After 24 hours, the cells were washed three times with phosphate-buffered saline (PBS) and cultured for the designated times. To assess the passage of infectivity, 12 mL of culture medium from the primary infected cells was collected, concentrated 100 times by filtration through Amicon Ultra-15, Ultracel-10K filters (Millipore, Carrigtwohill, Cork, Ireland), and 40 μL concentrated medium/well was used to infect naïve HuS-E/2 cells. All experiments were done with approval of the Ethical Committee of Kyoto University. Informed consent from patients was required for this approval.

Cloning and Sequencing.

To amplify the complementary DNA (cDNA) fragment corresponding to hypervariable region 1 (HVR-1),14 a nested RT-PCR was performed using Superscript III (Invitrogen, Carlsbad, CA) and PrimeSTAR HS DNA Polymerase (Takara, Tokyo, Japan). Reaction conditions were adjusted according to the manufacturer's protocol. Primers used were previously described15 and are shown in Supporting Table 1. PCR products were then purified and cloned using the Zero Blunt TOPO PCR Cloning Kit (Invitrogen). Ten recombinant clones were randomly isolated for each PCR product and sequenced as described.16

Quantitative Detection of HCV Core and Interferon alpha (IFN-α) Protein by Enzyme-Linked Immunosorbent Assay (ELISA).

The culture medium of infected cells was collected and concentrated 100 times as previously mentioned for the detection of HCV-core, or used directly for detection of IFN-α. HCV core protein was quantified using the Trak-C Core ELISA (Ortho Clinical Diagnostics, Neckargemünd, Germany). IFN-α was quantified using the Human IFN-A ELISA kit (PBL Biomedical Laboratories, Piscataway, NJ). Light absorbance was then measured using a Wallac 1420 multilabel counter (PerkinElmer Life Science, Waltham, MA).

Cytotoxicity Assay.

Culture medium was collected from HCV-infected cells and used for measuring lactate dehydrogenase (LDH) levels using an LDH cytotoxicity detection kit (Takara Biomedicals). Light absorbance was then measured as described above.

Sucrose Density Gradient.

The culture medium of the infected cells was collected, concentrated 500 times, and loaded onto a 20%-50% (wt/vol) sucrose gradient containing 50 mM PBS, 100 mM NaCl, and 1 mM EDTA, followed by centrifugation at 100,000g for 16 hours at 4°C in a SW41Ti rotor (Beckman, Fullerton, CA). The gradient was fractionated into 31 fractions that were used for HCV-RNA and core detection and HCV infection into naïve cells as described above.

Electron Microscopy.

The 1.12 g/mL fraction obtained by the sucrose density gradient showed the secondary infection activity as analyzed by transmission electron microscopy. The fraction was ultracentrifuged and the almost all supernatant was removed. The residual 10 μL of the solution was directly applied to a formvar-carbon grid for negative staining with 1% uranyl acetate solution and observed with an electron microscope (JEOL1010, JEOL, Tokyo, Japan).

Results

HuS-E/2 Cells Cultured in 3D/HF System Are Highly Permissive for Infection and Proliferation of bbHCV.

We compared the ability of HuS-E/2 cells cultured in the 3D/HF system (3D-HuS-E/2 cells) to those cultured as a monolayer (2D-HuS/E2 cells) to reproduce infection by HCV genotype 1b (HCV-RC6), derived from patient serum (RC6). The HCV-RC6 RNA levels in the 3D-HuS/E2 cells were significantly higher at all timepoints (Fig. 1A), showing that the 3D/HF system greatly improves the proliferation of bbHCV in HuS-E/2 cells. We observed that both the early stages of infection and the continuous replication of HCV-RC6 in HuS-E/2 cells was improved by 3D/HF culture when the culture conditions were changed after the infection from 3D/HF to 2D and vice versa (Supporting Fig. 2).

Figure 1.

Infection and proliferation of bbHCV in 3D-HuS-E/2 cells. (A) The quantity of HCV genomic RNA in 1 μg total RNA of 2D- or 3D-HuS-E/2 cells infected with HCV-RC6 was determined at each timepoint after infection by real-time RT-PCR analysis. (B) 3D-HuS-E/2 cells were infected with HCV-1b-containing sera: RC2, RC3, and P27; or HCV-2a-containing sera: 4: RC12, P17, P33. The quantity of HCV genomic RNA in the infected cells was determined as in (A). The culture medium from the last 2 days at each timepoint was collected, concentrated, and the amount of HCV-RNA (B) or HCV-core (C) was measured. Data represent the mean ± standard deviation (SD) of three independent experiments.

As reported,17 blocking CD81, an HCV-supposed entry receptor, during infection significantly impaired HCV proliferation into 3D-HuS-E/2 cells (Supporting Fig. 3), suggesting that CD81 is essential for HCV infectivity in 3D-HuS-E/2 cells. Although the expression level of CD81 mRNA in 3D-HuS-E/2 cells was observed, no significant change from 2D-HuS/E2 cells was found (data not shown), indicating that the quantity of CD81, at least, is not responsible for the improvement.

We then examined whether this system can be used for proliferation of six different bbHCV samples, three of which are HCV-1b (HCV-RC2, HCV-RC3, and HCV-P27) and three HCV-2a genotypes (HCV-RC12, HCV-P17, and HCV-P33) (Fig. 1B). Proliferation of HCV-RNA in the cells was seen in all six cases, suggesting that this system can be widely used for analysis of infection and proliferation of bbHCV strains. HCV-RNA and HCV-core were also detected in the culture medium (Fig. 1B). Different HCV strains showed variable patterns of proliferation and HCV-core secretion into the medium. Although HCV-core was detected from day 3 onward when RC3, RC6, and RC12 were used for infection, it was undetectable when RC2, P17, P27, and P33 sera were used, similar to 2D-HuS-E/2 cells infected with HCV-RC6 (Fig. 1C).

Production of Infectious Particles from 3D-HuS-E/2 Cells Infected with bbHCV.

The culture media from 2D or 3D-HuS/E2 cells infected with RC6 serum (Fig. 1A) were collected from days 5 to 7 postinfection (p.i.), concentrated, and inoculated into naïve 3D-HuS-E/2 cell culture media. HCV-RNA's proliferation in the infected cells was only detected when using the culture medium from 3D-HuS-E/2 cells and not 2D-HuS-E/2 cells (Fig. 2A). Media collected from HCV-RC3 at days 5 to 7 and from HCV-RC12 from days 23 to 25 p.i. were also able to infect naïve cells (Fig. 2A). These data suggested the production and secretion of infectious virus-like particles. To investigate this further, biophysical analysis was performed. The culture medium of HCV-RC3 infected 3D-HuS-E/2 cells at day 7 p.i. was fractionated using a sucrose density gradient after RNase A treatment. HCV core was detected in the 1.11 to 1.14 g/mL fractions; similarly, the nuclease-resistant HCV RNA peaked in the 1.12 g/mL fraction (Fig. 2B). Furthermore, only the 1.12 g/mL fraction was able to infect naïve cells as examined above (data not shown). This fraction was pelleted by ultracentrifugation and examined by electron microscopy with negative staining. We observed 33-nm to 45-nm diameter spherical particles (Fig. 2C, panel 1) with spike-like structures from 7-9 nm in length on the surface (Fig. 2C, panel 2), consistent with HCV morphology reported previously in HCV patients.18 These were detected in the sample collected from HCV-RC3-treated but not mock-treated 3D-HuS-E/2 cells. These data suggest that production of infectious virus-like particles occurs in 3D-HuS-E/2 cells infected with some bbHCV strains. It is therefore likely that 3D-HuS-E/2 cells can be used to reproduce nearly all steps in the HCV life cycle.

Figure 2.

Production of infectious virus-like particles from 3D-HuS-E/2 cells infected with different HCV strains. (A) The culture medium of 3D-HuS-E/2 cells infected with HCV-RC3 or HCV-RC6 was collected from days 5 to 7 p.i. and for HCV-RC12 from days 23 to 25 p.i. The culture medium of 2D-HuS-E/2 cells infected with HCV-RC6 was also collected from days 5 to 7 p.i., and used to treat naïve 3D-HuS-E/2 cells. The quantity of HCV genomic RNA in 1 μg of total cellular RNA was determined as in Fig. 1. (B) The concentrated culture medium of 3D-HuS-E/2 cells infected with HCV-RC3 was collected from days 5 to 7 p.i., and fractionated by ultracentrifugation with a 20%-50% sucrose density gradient. HCV-core protein and the RNase A-resistant HCV-RNA in the different fractions were quantitatively analyzed using an HCV-core ELISA kit and real-time RT-PCR, respectively. Data represent the mean ± SD of three independent experiments. (C) Photomicrograph showing negatively stained virus-like particles from the culture medium of HCV-RC3-infected 3D-HuS-E/2 cells (arrowheads, panels 1 and 2). The arrows indicate the spike-like structures found on the surface of the virus-like particles (panel 2).

Prolonged Culture of HCV-Infected Cells in the 3D Hollow Fiber System.

For HCV-RC6-infected cells (Fig. 3A), the amount of HCV-RNA in the cells fluctuated during the 30-day culture period. The levels of both HCV-RNA and HCV-core in the medium showed a similar pattern of fluctuations that peaked on days 5 and 20 p.i. Unlike RC6, the pattern of HCV-RNA levels in the medium of RC12-infected cells showed a negative correlation with that detected in the cells. This was clearly seen on day 10 p.i., when a sharp increase and decrease of HCV-RNA in the medium and the cells, respectively, was observed (Fig. 3B). Similarly, the amount of HCV-core detected in the medium throughout the culture was not correlated with RNA levels in the medium. Instead, core levels were very low in the first 10 days, at which time levels increased, reaching a peak on day 20 p.i. (Fig. 3B). Culture media from cells infected with HCV-RC6 from days 5 to 7 and 20 to 30 p.i. (Fig. 3A) and that from HCV-RC12 from days 20 to 25 p.i. showed passage of infectivity (Fig. 3B). All culture media showing infectivity appeared to have a high amount of HCV-core protein.

Figure 3.

Prolonged culture of HCV-infected cells in the 3D/HF system. After infection with HCV-RC6 (A) and HCV-RC12 (B), 3D-HuS-E/2 cells were cultured for 30 days with a medium change every 2 days. The HCV-RNA in the cells and medium as well as the HCV-core in the medium were quantitatively analyzed at the designated timepoints as in Fig. 1. Culture media were also used to treat naïve 3D-HuS-E/2 cells to examine the secondary infection as in Fig. 2. (+) and (−) indicate detection or no detection of secondary infection. (C) Culture media of HCV-RC6 and HCV-RC12 infected cells collected at each timepoint were used for the detection of LDH levels released from dead cells. LDH levels were normalized to uninfected cells cultured for the same time. Data represent the mean ± SD of three independent experiments.

Clonal Changes in HCV During Prolonged Culture.

In order to perform a populational analysis to understand the fluctuating pattern seen during HCV proliferation, two sera with limited HCV variants, HCV-RC6 (two major strains) and -RC12 (single major strain) from immunosuppressed liver transplantation patients with recurrent HCV were used in the previous prolonged infection experiment. The variants' composition was analyzed by single-strand confirmation polymorphism analysis for HCV-HVR1 (Supporting Fig. 4). RC6 serum (Fig. 4A) showed two different major sequences, HCV-RC6-1 and -2 strains, which constituted 60% and 40%, respectively, and shared 85% homology. In cells infected with HCV-RC6 the nucleotide sequence of HVR1 on day 5 showed 97% homology to HCV-RC6-1, and on day 20 p.i. it showed 97% homology to HCV-RC6-2. These data suggest selection of the dominant HCV strain in the cells over time. For RC12 (Fig. 4B), the nucleotide sequence on day 5 p.i. had only one nucleotide difference from that of the HCV from the original serum. The sequence from day 20 p.i. was four nucleotides different from that from the serum, and five different from the cells on day 5 p.i. These data indicated that each peak of HCV-RNA that appeared in the cells infected with RC12 serum included primarily a single HCV strain with a slightly different genomic sequence. This suggests that the periodic appearance of HCV-RNA peaks in the cells infected with a particular HCV strain is a result of selection and/or mutation of HCV strains during the prolonged culture period.

Figure 4.

Comparison of HCV-HVR1 sequences in the serum used for infection and the HCV replicating in the cells on days 5 and 20 after infection of HCV-RC6 (A) or HCV-RC12 (B). Nucleotide numbering was based on HCV-J1 sequence (GenBank Access. No. D10749). Three additional nucleotides were found at the 5′-terminal end of the E2 regions of all RC6 sequences. The major sequence present in the serum used for infection is shown in the upper row in each panel. Dots represent the identical nucleotides.

Cellular Response Induced by bbHCV Infection.

At day 10 p.i., HCV-RNA levels in the culture medium rose and RNA levels in 3D-HuS-E/2 cells infected with HCV-RC12 dropped (Figs. 1B, 3B). To determine if this was caused by a cytotoxic effect of HCV infection, LDH levels were measured in the culture medium of HCV-RC6- and HCV-RC12-infected 3D-HuS-E/2 cells. LDH activity showed a strong correlation with HCV-RNA levels in the medium on day 10 p.i. in HCV-R12-infected cells (Fig. 3B), suggesting a cytotoxic effect of HCV-RC12 that was not observed in the case of HCV-RC6 (Fig. 3A,C). To determine if this HCV infection-mediated cytotoxicity is due to apoptosis, as with other viruses belonging to the Flaviviridae family,19 the involvement of caspase was examined using the caspase inhibitor z-VAD-fmk. A significant dose-dependent reduction in HCV-RNA levels in the medium and LDH activity (Fig. 5A,B) was found, whereas no significant effect was observed on the viability of noninfected cells (Fig. 5B) or intracellular HCV-RNA levels (Fig. 5A). This suggested that the cytotoxic effect of HCV infection is mediated by apoptosis. It is noteworthy that HCV-induced cytopathicity was also found when HCV-P17 and HCV-P33 samples were used for infection (both are HCV-2a genotype) and was not reproduced in any of the HCV-1b genotype samples used in this work (Fig. 5C).

Figure 5.

Cellular response of 3D-HuS-E/2 cells infected with bbHCV. 3D-HuS-E/2 cells infected with HCV-RC12 and mock-treated cells were cultured for 10 days in the presence of z-VAD-fmk (0, 10, and 20 μM). (A) HCV-RNA in the cells and medium on day 10 was measured as in Fig. 1. (B) LDH levels in the medium on day 10 after infection with HCV-RC12 was measured as in Fig. 3. (C) Culture media of HCV-RC3, HCV-P17, HCV-P27, HCV-P33, and mock-infected cells collected at designated points were used for the detection of LDH levels. (D) IFN-α levels in the culture media of HCV-RC6, HCV-RC3, and mock-infected cells collected at each designated timepoint were measured by ELISA. Data represent the mean ± SD of three independent experiments.

After infection with HCV-RC6, no cytotoxicity was detected that might have inhibited HCV-RC6-1 proliferation in the cells. However, HCV-RC6-2 RNA replaced HCV-RC6-1 RNA during prolonged culture. To assess a possible role of the innate-immune response in this phenomenon, the production of IFN-α in the medium was measured during the first 11 days p.i. IFN-α production was detected as early as day 1 p.i., reached a peak at day 7 p.i., and was then rapidly lost (Fig. 5D). These data suggest that HCV-RC6-1 infection induced the innate-immune response of the cells, possibly leading to suppression of its proliferation. In contrast to HCV-RC6-1, HCV-RC3 did not show any stimulation of IFN-α production upon infection in the first 10 days, showing a possible strain-dependent evasion from the host defense within the same genotype.

Discussion

In this study we report the development of a novel system that reproduces bbHCV infection, proliferation, and production of infectious virus. The most recent models used in the study of the life cycle of HCV infection are based on subclones of HuH-7 cells infected with JFH1 recombinant virus or its derivatives.4 HuH-7 cells and its subclones, however, do not support the entire life cycle of the bbHCVs present in patients' blood.5 Moreover, HCV has considerable diversity and variability. It is generally classified into six major genotypes and more than 100 subtypes.20 This huge pool of natural HCV variants causes a wide variety of diseases, including chronic hepatitis, cirrhosis, and hepatocellular carcinoma.21 JFH1, however, is a single isolate of HCV genotype 2a that was originally derived from a patient with rare fulminant hepatitis.4 We suggest that our newly established system has an important advantage because it supports the entire life cycle of a variety of HCV strains and genotypes.

Due to the lack of some in vivo factors, including host immune response, in vitro systems may not completely reproduce the in vivo situation. However, in vitro experimental systems seem to be important to simplify particular events from the complex situation in vivo. From that standpoint, our cell culture system is likely reproducing the early event of HCV infection in the absence of host-immune responses and supporting whole life cycle of the blood-borne HCV. Several in vitro hepatocyte culture systems have been reported to be useful for studying the infection and replication of bbHCV.5–8, 22 Only the radial-flow bioreactor (RFB) 3D culture system demonstrated production of infectious viruses.22 In our studies we observed not only the enhancement of HCV replication, but also the production of infectious HCV particles in the medium using the 3D/HF system. These data suggest that some structure of the cell mass formed by the 3D culture system, most likely the polar character, is essential for the life cycle of bbHCV. The RFB system is composed of a dedicated device containing 1 × 109 FLC4 cells with a culture area of 2.7 m2.22 It can only be used to study HCV particle production in the medium and not the cellular events that accompany the HCV life cycle. In contrast, because cells grown in our 3D/HF system are cultured in 12-well plates at a density of 3 × 105/fiber, it is much simpler to study both viral and cellular events.

The production of infectious particles was not detected with infection by different HCV strains, despite detecting equivalent levels of HCV-RNA in the cells (Fig. 1B,C). Delayed production of infectious particles was also observed in cells infected with HCV-RC12 after prolonged culture. A similar delay was also observed in the RFB system.22 Considering the relative stability of HuS-E/2 cells5 and the relatively high frequency of the change in HCV population in the cells,16 it is likely that mutation of the HCV genome and/or selection of clones during prolonged culture improved the productivity of infectious particles. A marked improvement of infectious particle production by substitution of the structural proteins of the genome was also reported in the recombinant HCV production system.23 The lack of production of infectious particles soon after infection may serve to avoid an early strong response from the host immune system, and demonstrates a novel mechanism of latent infection by HCV. Although they may not be associated with plasma components as those present in vivo, HCV virus-like particles produced by this system showed a close resemblance to those isolated from infected HCV patients because they showed the same size18 and were within the fraction range.24 They may help in the study of viral and cellular factors required for particle production and the possible receptors utilized for infection with different HCV strains.

Fluctuation in HCV proliferation was observed during the prolonged culture of 3D-HuS-E/2 cells infected with bbHCV (Fig. 3A,B), consistent with previous reports in other culture systems.6, 22 This fluctuation was associated with a change in viral quasispecies, suggesting that an HCV strain having a growth advantage proliferates selectively and dominantly in these culture conditions. Because the progressive emergence of each dominant strain was only temporary, it is highly likely that the infection and proliferation of such an HCV strain is suppressed by cellular mechanism(s). Our results suggest that there are actually two cellular mechanisms functioning to do this. The first is the involvement of the innate immune system, as evidenced by the secretion of IFN-α during the first week of infection (Fig. 5D). This is the first report of secretion of IFN-α from cultured cells infected with bbHCV. Although recent reports suggest that stimulation of the IFN pathway by HCV infection could be impaired by HCV NS3-4a proteinase-mediated cleavage of IPS-1,25 our results suggest that not all bbHCVs possess a host cell suppressive function. The second mechanism is HCV-induced cell death (Fig. 3C). Almost all the studies reporting HCV-induced apoptosis used hepatocellular carcinoma cell lines.26, 27 Because it has been established that the inability to undergo apoptosis is essential for the development of cancer,28–30 our use of immortalized, noncancerous HuS-E/2 hepatocytes may make it possible to reproduce the physiological response of the cells to bbHCV infection more closely. Although HCV-induced apoptosis was not found when HCV-1b was used for infection, it was found in all cases where HCV-2a was used, suggesting a higher cytopathic tendency of the HCV-2a genotype. HCV proliferation was continuously found even after the suppression of the first peak of RNA production during prolonged culture. How HCV survives under those conditions is still unknown. Further studies to clarify the molecular mechanisms involving the HCV-cell interaction can be done using this novel 3D culture system that reproduces the infection of a variety of bbHCVs.

In conclusion, we have established a new in vitro culture system that can support the entire life cycle of a variety of HCV isolates and genotypes. Although this in vitro model system may not completely reproduce the in vivo situation, we believe it is the first in vitro system showing HCV strain-dependent virus/cell interaction including induction of cellular apoptosis and/or evasion from cellular innate immune response, which may make it a good tool for analysis of virus/host interaction together with the development of new anti-HCV strategies for the different bbHCV strains.

Acknowledgements

We thank T. Yamaguchi for providing hollow fibers and culture medium.

Ancillary