Unexpected host range of hepatitis C virus replicons


Zhu Q, Guo J-T, Seeger C. Replications of hepatitis C virus subgenomes in nonhepatic epithelial and mouse hepatoma cells. J Virol 2003; 77: 92049210. (Reprinted with permission from the American Society for Microbiology (ASM).)

Ralf Bartenschlager*, * Department of Molecular Virology University of Heidelberg Heidelberg, Germany.


The hepatitis C virus (HCV) pandemic affects the health of more than 170 million people and is the major indication for orthotopic liver transplantations. Although the human liver is the primary site for HCV replication, it is not known if extrahepatic tissues are also infected by the virus nor if nonprimate cells are permissive for RNA replication. Because HCV exists as a quasispecies, it is conceivable that a viral population may include variants that can replicate in different cell types and other species. We have tested this hypothesis and found that subgenomic HCV RNA can replicate in mouse hepatoma and nonhepatic human epithelial cells. Replicons isolated from these cell lines carry new mutations that could be involved in the control of tropism of the virus. Our results demonstrated that translation and RNA-directed RNA replication of HCV do not depend on hepatocyte or primate-specific factors. Moreover, our results could open the path for the development of animal models for HCV infection.


Viruses are obligate intracellular parasites that absolutely depend on a functional host cell machinery for their multiplication. Consequently, studies of the viral life cycle require systems that support virus propagation in cultured cells; unfortunately, these systems are not always available. The hepatitis C virus (HCV) is a prominent example how the lack of a suitable cell culture system can impede scientific progress for many years.1, 2 A breakthrough was the development of the so-called “replicon system,” which is based on the autonomous replication of HCV subgenomes in the human hepatoma cell line Huh-7.3 These replicons are composed of (1) the HCV 5′ nontranslated region, (2) the selectable marker neomycin phosphotransferase, which confers resistance against the cytotoxic drug G418, (3) a heterologous internal ribosome entry site (IRES) that allows efficient expression of the HCV NS3 to NS5B proteins, and (4) the HCV 3′ nontranslated region (Fig. 1). Upon transfection of Huh-7 cells with replicon RNA molecules and subsequent selection, G418-resistant cell clones could be established that carry large amounts of HCV RNA and express substantial levels of viral proteins. Studies performed in the past few years have revealed two major determinants for efficient RNA replication in Huh-7 cells: cell culture adaptive mutations in the HCV coding region of the replicons and host cell permissiveness.4–14

Figure 1.

Comparison of the HCV genome organization (upper panel) with the structure of a subgenomic selectable replicon (lower panel). The 5′ and 3′ nontranslated regions are indicated with small bars and sequences encoding the neomycin phosphotransferase (neo) or the HCV polyprotein with thick bars. Translation of the neo gene is directed by the HCV IRES, whereas the proteins required for RNA replication (NS3 to NS5B) are translated under control of the IRES from another virus (the encephalomyocarditis virus). Functions ascribed to HCV proteins that are required for RNA replication are indicated below the replicon scheme. The positions of amino acid substitutions mentioned in the text are indicated above. The role of NS5A in RNA replication is as yet unknown. C, core protein; E1 and E2, envelope glycoproteins; NTPase, nucleoside triphosphatase.

Most adaptive mutations have been identified in NS3, NS4B, and NS5A and have been shown to cluster in certain regions of these proteins. The most prominent hot spots reside at two distinct positions in NS4B and in the center of NS5A, where serine residues that are involved in phosphorylation of this protein are often replaced by nonphosphorylatable amino acid residues.4, 9 The most potent effect on RNA replication is mediated by combinations of mutations in NS3 (e.g., E1202G) with single substitutions in NS4B or NS5A (e.g., S2204I).9, 15 The second determinant influencing HCV RNA replication is the permissiveness of the host cell; that is, the extent to which a given cell supports replication of this RNA. Only a low number of cells in a culture appear to be sufficiently permissive, and these cells are enriched during G418 selection. This conclusion is based on the observation that removal of the replicon from replicon-harboring cells (e.g., by treatment with interferon-α [IFN-α] or a selective inhibitor) results in a population of “cured” cells that support higher levels of HCV RNA replication compared with naïve cells.7–9

Since the original description in 1999, the replicon system has been improved substantially. Transient replication assays as well as full length genomes replicating transiently and stably in Huh-7 cells have become available.11, 12 However, two major limitations existed: the lack of virion production and the restriction of HCV replicons to Huh-7 cells. The latter limitation has now been overcome by Zhu and coworkers, who reported the successful propagation of HCV replicons in HeLa cells and the mouse hepatoma cell line Hepa1-6.16 Their experimental approach was based on the assumption that the replication-enhancing mutations observed in Huh-7 cells (e.g., S2204I) reflect an adaptation to a particular host cell environment and that further mutations might be required to adapt replicons to other cell lines. Thus, when using complex mixtures of replicon RNAs with a high genetic heterogeneity, the chances to select for an adapted variant should be much higher compared with the use of rather uniform replicon RNA that is generated by in vitro transcription from cloned HCV templates.

Zhu and coworkers reasoned that because of the high error rate of the viral replicase, HCV replicons exist within a single cell as a swarm of sequence variants in which (on a statistical basis) each RNA molecule differs from the other at least at one position. Therefore, they isolated total RNA from Huh-7 cell clones that carried stably replicating HCV replicons. These total RNA preparations were used to screen various cell lines, including BHK (baby hamster kidney cells), Vero (kidney epithelial cells from the African green monkey), HeLa (human cervix carcinoma cells), and Hepa1-6 (hepatoma cells from the mouse) by means of RNA transfection and subsequent G418 selection. Only in HeLa cells was a low number of G418-resistant colonies obtained. Analysis of seven stable cell clones revealed that 90% of the cells expressed viral proteins at a given time point, with replicon RNA copy numbers in the range of 3,000 molecules per cell.

Although these properties are comparable to Huh-7 replicon cell clones, distinct differences have been described in this and a subsequent study by the same group.17 First, although replicon RNA copy numbers in Huh-7 cells very much depend on host cell growth, this does not appear to be the case for HeLa cells. This observation argues for some fundamental differences in host cell requirements of HCV replicons in HeLa versus Huh-7 cells; however, the discrepancy could also be explained by a much less pronounced inhibition of HeLa cell growth by cell-to-cell contact. Second, the IC50 observed for IFN-α was approximately tenfold lower in HeLa cells compared with Huh-7 cells (0.1 IU/ml vs. ≈1 IU/ml, respectively).17 Third, treatment of HeLa cells carrying HCV replicons with IFN-α led to apoptosis in more than 30% of cells. However, the induction of cell death could be prevented when replicons were blocked with an NS5B-specific inhibitor prior to addition of IFN-α, suggesting that viral RNA or protein stimulated an innate cellular response that sensitized cells towards apoptosis.17 In contrast, Huh-7 replicon cells treated in a similar fashion did not undergo programmed cell death.

To determine whether or not HCV replicons in HeLa cells had acquired cell type–specific adaptive mutations, Zhu and coworkers compared the number of G418-resistant colonies obtained after transfection of HeLa cells with total RNAs isolated from HeLa or Huh-7 cells harboring selectable replicons.16 As summarized in Table 1, replicon RNA isolated from HeLa cells yielded significantly more colonies upon transfection of naïve HeLa cells compared with naïve Huh-7 cells. In contrast, replicon RNAs isolated from Huh-7 cells were more efficient when introduced into naïve Huh-7 cells as compared with naïve HeLa cells. Because the number of G418-resistant colonies is a measure of the efficiency of RNA replication, these data indicated the selection for cell type–specific replicon variants that multiplied more efficiently in HeLa cells. Nucleotide sequence analysis revealed that the major cell culture adaptive mutations present in Huh-7 cells (E1202G and S2204I) were maintained in replicons isolated from HeLa cells, but several additional mutations were found that so far had not been observed in Huh-7 cells. Most notable was an amino acid substitution in the N-terminus of NS4B (V1749A), because it was conserved in replicon RNAs isolated from two independent HeLa replicon cell clones.

Table 1. Increase of the Number of G418-Resistant Replicon Cell Clones by Prior Passage of Replicon RNA in the Same Cell Line
Source of Replicon RNATransfected Into
  • *

    Number of G418-resistant colonies obtained per ng replicon RNA that was present in total RNA of a given replicon-harboring cell clone and used for transfection. Data are taken from Table 2 of Zhu and coworkers.16


Encouraged by these observations, Zhu and coworkers investigated whether or not this or other additional mutations would expand the tropism of such replicons beyond Huh-7 and HeLa cells. A panel of 10 different cell lines of human, monkey, rat, hamster, and mouse origin was transfected with total RNA isolated from HeLa replicon cells. However, transfection of only the mouse cell line Hepa1-6 yielded G418-resistant colonies. Sequence analysis revealed that replicons isolated from these mouse cells had preserved the majority of mutations found in HeLa cells that were used as the replicon RNA source, and only a few additional mutations were detected.

In an attempt to generate cloned versions of cell type–adapted replicons, a panel of selectable replicon constructs was made that carried various combinations of mutations identified by sequence analysis of HeLa cell–derived replicons. Two of these constructs yielded a low number of G418-resistant colonies in HeLa cells but not in Hepa1-6 cells. Interestingly, in both cases the replicons contained the novel NS4B mutation (V1749A) in different combinations with other mutations. Unfortunately, no replicon was tested that carried this NS4B mutation alone or only in combination with the highly adaptive ones (E1202G and S2204I). Nevertheless, these data suggest that the NS4B mutation contributes to the expanded tissue tropism. However, the fact that no colonies were obtained in the mouse liver cell line Hepa1-6 indicates that this mutation is HeLa cell–specific.

Given the low number of colonies obtained with HeLa cells and the negative results obtained with Hepa1-6 cells, Zhu and coworkers wondered if they had missed replicon variants that carry more efficient cell type–specific adaptive mutations but represent a very minor species in the replicon RNA population present in these cells. Therefore, they constructed replicon complementary DNA libraries derived from Huh-7, HeLa, and Hepa1-6 replicon cells and transfected replicon RNA pools with a complexity of 2,000 variants into the different cell lines. This approach enabled them to establish replicon RNA-containing clones of Hepa1-6 cells. It will be interesting to see which replicon variant was able to establish sustained RNA replication in these mouse cells.

In conclusion, the study by Zhu and coworkers represents an important extension of the HCV replicon system. Although transient replication assays are not yet possible with HeLa and Hepa1-6 cells, and the number of G418-resistant colonies obtained even with cell culture–adapted replicons is low, the availability of such alternative replicon cell clones is critical for a better understanding of the HCV life cycle. Because cell lines can deviate quite substantially in their morphologic and physiologic state from cells in vivo, the availability of multiple replicon cell lines avoids possible pitfalls that can be encountered when working with only one particular cell line. Moreover, this study clearly shows that HCV RNA replication does not depend on liver cell–specific factors but rather can replicate in nonliver tissues. This observation corroborates earlier studies describing low-level HCV replication in infected nonhepatic cell lines like Daudi (B cell) or Molt-4 and MT-2 (T cell). Finally, the finding that HCV replicons can replicate in mouse cells and that replication in such nonhuman cells may be achieved by certain adaptive mutations offers some hope for the development of a mouse model for HCV infection. Although this is probably still a long way to go, the first step has been made.