Replication of a hepatitis C virus replicon clone in mouse cells

Background Hepatitis C Virus (HCV) is a significant public health burden and small animal models are needed to study the pathology and immunobiology of the virus. In effort to develop experimental HCV mouse models, we screened a panel of HCV replicons to identify clones capable of replicating in mouse hepatocytes. Results We report the establishment of stable HCV replication in mouse hepatocyte and fibroblast cell lines using replicons derived from the JFH-1 genotype 2a consensus sequence. Viral RNA replication efficiency in mouse cells was comparable to that observed in human Huh-7 replicon cells, with negative-strand HCV RNA and the viral NS5A protein being readily detected by Northern and Western Blot analysis, respectively. Although HCV replication was established in the absence of adaptive mutations that might otherwise compromise the in vitro infectivity of the JFH-1 clone, no infectious virus was detected when the culture medium from full length HCV RNA replicating mouse cells was titrated on Huh-7 cells, suggesting that the mouse cells were unable to support production of infectious progeny viral particles. Consistent with an additional block in viral entry, infectious JFH-1 particles produced in Huh-7 cells were not able to establish detectable HCV RNA replication in naïve mouse cells. Conclusion Thus, this report expands the repertoire of HCV replication systems and possibly represents a step toward developing mouse models of HCV replication, but it also highlights that other species restrictions might continue to make the development of a purely murine HCV infectious model challenging.


Background
Hepatitis C virus (HCV) is an enveloped, positive-strand RNA virus that causes acute and chronic hepatitis [1]. Between 70-90% of those who become infected fail to clear the virus and remain chronically infected with the risk of developing liver cirrhosis and hepatocellular carcinoma [2]. Unfortunately, there is no vaccine available to prevent this infection, and the only approved treatment has toxic side effects and is only effective in a subset of patients [3,4].
Even though recent work has led to the development of in vitro HCV infection systems, which allow for molecular analysis of the entire viral life cycle [5][6][7], the study of the immunobiology and pathogenesis of HCV still requires the development of genetically defined small animal models. One obstacle to the development of HCV mouse models has been the limited host range of the virus. The restrictions that block HCV infection in mice are not well defined, but appear to involve multiple steps such as viral entry and genome replication.
Notably however, HCV replicons based on engineered viral genomes into which the antibiotic resistant marker neomycin phosphotransferase (neo) has been inserted [8] (Fig. 1A) provide a means of experimentally by-passing viral entry and actively selecting for HCV replication after transfection of RNA into cells. The ability to select for cells replicating the neo-expressing replicon RNA led to the discovery that efficient replication of most HCV replicons in cell culture requires adaptive mutations in the viral genome [9][10][11][12][13]. Although HCV replication initially could only be achieved in the human hepatoma cell line, Huh-7, the ability to select for replication enhancing mutations eventually led to the establishment of HCV replication in other hepatic (HepG2 and IMY-N9 [14]) and nonhepatic (HeLa [15,16] and HEK293 [15,17]) human cell lines.
Unlike other published studies that focus exclusively on HCV replication in human and/or primate cell lines, Zhu et al (2003) further demonstrated that replication of the HCV-N genotype 1b subgenomic replicon could be initiated in one of the several mouse cell lines tested. However, this replication could only be established in a single mouse cell line after transfection of total RNA extracted from HeLa cells that were already replicating the adapted replicon (i.e. total human cellular RNA presumably containing a quasispecies of HCV replicons) [16]. In contrast, HCV replication could not be initiated in these mouse cells by transfection of in vitro transcribed replicon RNA generated from either the parental replicon construct or from any of the "adapted" replicon clones isolated from their original mouse replicon cells. Hence, no "mousepermissive" HCV replicon clone was identified.
Because the development of HCV mouse models would be greatly facilitated by the identification of defined HCV clone(s) capable of establishing and maintaining replication in mice, we assembled a panel of HCV replicons derived from different HCV genotypes and assessed their ability to replicate in mouse cells. We show that JFH-1 genotype 2a subgenomic and full length replicons are able to stably replicate in multiple mouse hepatocyte and fibroblast cell lines following transfection of the replicon RNA. Although the HCV replication achieved in mouse hepatocytes was not dependent on adaptive mutations that might compromise the infectivity of the viral clone, no infectious HCV was detected in the media of mouse cells replicating full length HCV RNAs, nor were infectious JFH-1 virus particles produced in Huh-7 cells able to infect naïve mouse cells. Thus, these mouse cells are permissive for JFH-1 HCV replication, but exhibit blocks in both viral entry and production of infectious particles.

Establishment of G418-resistant mouse hepatocyte colonies after transfection of subgenomic replicons encoding the neomycin selection gene
Because different HCV replicon clones exhibit a range of replication efficiencies in Huh-7 cells, we hypothesized that different replicons might also display differences in their ability to replicate in mouse cells. Thus, we screened a panel of subgenomic HCV replicons derived from genotypes 1b, 1a, and 2a for their ability to confer G418 resistance to mouse hepatocytes in vitro. We initially screened for HCV replication in immortalized Met Mouse Hepatocytes (MMH) cells; however, because HCV replication efficiency within different Huh-7 cell lines can vary, we tested (A) Schematic diagram of HCV genomic and replicon RNA HCV replication in two independently derived MMH cell lines, MMHD3 [18] and MMH1-1 [19].
Replicon RNA from the clones listed in Table 1 was transcribed in vitro and electroporated into MMH cells. Transfected cells were then treated with 500 μg/ml G418, which was the minimal concentration required to effectively kill untransfected control cell cultures. Despite the fact that G418-resistent Huh-7 cell colonies survived the selection process after transfection with all of the replicon RNAs tested, RNA transcribed from the HCV genotype 1b clones (Con1 and HCV-N) and the HCV genotype 1a H77 clone did not generate G418-resistent mouse cell colonies (data not shown). In contrast, transfection of genotype 2a subgenomic (sg) JFH-1 RNA produce G418-resistant mouse hepatocyte colonies in both MMH cell lines in multiple independent transfection experiments (Fig. 1B).

Expression and replication of sgJFH-1 HCV replicons in immortalized mouse hepatocytes
To confirm the expression and replication of HCV in MMHD3 and MMH1-1 cells, G418-resistant cell colonies were expanded for further analysis. Western Blot analysis of total cell lysate verified that the viral NS5A protein was present in each of the MMHD3 and MMH1-1 replicon cell clones ( Fig. 2A, lanes 5, and 6-11, respectively). Yet, constitutive HCV protein expression was not due the integration of an HCV transgene, as PCR analysis on cellular genomic DNA revealed no evidence of JFH-1 sequences (data not shown).
RT-QPCR analysis of the same RNA samples indicated that HCV RNA levels among the mouse sgJFH-1 replicon clones ranged from 2.9 × 10 6 -7.0 × 10 7 copies per microgram (copies/μg) of total cellular RNA. This was similar to the 7.4 × 10 6 -8.5 × 10 7 copies/μg observed in the human sgJFH-1 replicon clones. Thus, the average 2.0 × 10 7 HCV RNA copies/μg calculated in the mouse cells was comparable to the average 5.0 × 10 7 HCV RNA copies/μg detected in the Huh-7 clones indicating that while the level of the HCV RNA does vary between individual cells clones, no statistical difference in RNA level was detectable between mouse and human cells.

Adaptive mutations are not required for replication of sgJFH-1 replicons in mouse hepatocytes
Because adaptive mutations are required for efficient replication of many HCV replicons in cell culture [9][10][11][20][21][22], we sequenced replicon clones obtained from our mouse cell lines to determine whether the population of sgJFH-1 replicons present in the MMH cells had acquired any adaptive mutations. RNA was extracted from 3 of the MMH replicon cell lines (MMHD3-sgJFH#1 and MMH1-1sgJFH#4 and #9). After reverse transcription, 4 PCR primer sets were used to amplify overlapping segments spanning the sgJFH-1 cDNA (Fig. 3A). These PCR products were ligated into the pGEMT-Easy vector and 12 individual clones of each fragment were sequenced. Although random mutations were observed throughout the cloned genomes, none of the changes were detected in more than 1 of the 12 clones sequenced (Fig. 3B), supporting the conclusion that adaptive mutations had not become established in the mouse replicon population. Also consistent with the absence of replicon adaptive mutations, total RNA isolated from both MMH and Huh-7 replicon cells containing comparable copies of sgJFH RNA (as determined by RT-QPCR) formed equivalent numbers of G418-resistant colonies/μg of viral RNA when re-transfected into naïve MMH cells (data not shown).
To determine if the initial establishment of the mouse replicon clones had selected for adapted MMH cells that were more permissive for HCV replication, we also "cured" the sgJFH-1 replicon from our MMH replicon cells with a 3 week treatment of IFNα (1000 U/ml) and IFNγ (500 U/ ml), but subsequent sgJFH-1 RNA transfections into these "cured" (e.g. G418-sensitive, HCV-negative) mouse cells did not produce higher numbers of G418-resistent colonies/μg of input RNA compared to parallel transfections into the parental MMH cell lines indicating that we had not selected MMH cells that more efficiently allow for establishment of HCV replication (data not shown).

Replication of sgJFH-1 HCV in other mouse cell lines
Because the sgJFH-1 replicon was able to replicate in MMH cells without any apparent viral or cellular adaptive mutations, we proceeded to determine if this ability was restricted to MMH cells or if the JFH-1 clone could replicate in other mouse cell lines. Hence, we transfected in vitro transcribed sgJFH-1 RNA into mouse AML12 hepatocytes as well as mouse NIH3T3 embryonic fibroblasts and selected for cells that were able to support replication of the neo-expressing replicon. Figure 4A shows strand-specific HCV Northern Blot analysis that verifies the presence of both positive-strand (Fig. 4A, top panel) and negativestrand (Fig. 4A, middle panel) HCV RNA in all the AML12 mouse hepatocyte cell clones. Similarly, negative strandspecific HCV Northern Blot analysis of total RNA from 9 NIH3T3 replicon cell clones also confirmed the presence of negative-strand HCV RNA in these non-hepatic mouse cells demonstrating that the sgJFH-1 HCV clone can replicate in a variety of mouse cell lines (Fig. 4B).

Lack of infectious HCV particle entry into and secretion from mouse hepatocytes
Because the flJFH-1 is capable of producing infectious HCV particles, we proceeded to determine whether the MMH1-1 mouse cells were able to secrete infectious HCV particles and/or were permissive for flJFH-1 entry. To determine if the flJFH-1 replicon clones were secreting infectious HCV particles that could transmit G418-resistance, the culture media from individual clones was collected and used to inoculate naive Huh-7 and MMH1-1 cells. Although G418-resistant colonies formed when Huh-7 cells were inoculated with the supernatant from Huh-7 flJFH-1 replicon cells, the culture supernatant from MMH1-1 flJFH-1 replicon cells did not confer G418resistance to Huh-7 cells suggesting that the mouse cells were not able to secrete infectious HCV particles. Titration of the same media on Huh-7 cells further confirmed that all the Huh-7 flJFH-1 cell clones were secreting infectious HCV particles (expressed as foci forming units per ml), while no infectious HCV was detected in the media from MMH1-1 flJFH-1 cells, even after 50-fold concentration by centrifugal filtration (Fig. 5B).
Notably, inoculation of naïve MMH1-1 cells with media from either Huh-7 flJFH-1 or MMH1-1 flJFH-1 cell cultures did not confer G418-resistence to the mouse cells confirming that HCV particles are unable to enter mouse cells. Importantly, inoculation with high titer authentic HCV protein and RNA detection in immortalized mouse hepatocytes HCV virus produced in Huh-7 cells also failed to establish JFH-1 replication in MMH1-1, AML12, or NIH3T3 mouse cells (data not shown).
To determine whether the lack of detectable infectivity in the culture medium of flJFH-1 MMH1-1 replicon clones was due to an inability to secrete HCV particles or due to the secretion of non-infectious HCV particles, the culture supernatant from Huh-flJFH#7 and MMH1-1-flJFH#2 cultures were subjected to sucrose density gradient centrifugation. Fractions were collected, total RNA extracted, and RT-QPCR analysis performed to determine the amount of viral RNA in each fraction (Fig. 5C). HCV RNA in the supernatant of the Huh-flJFH#7 cells was detected over a broad density range of 1-1.15 g/ml, with a predominant peak at 1.07 g/ml, supporting the conclusion that the viral RNA was associated with enveloped particles (Fig. 5C). This is similar to the HCV RNA density profile reported for in vitro-derived infectious JFH-1 particles from which RNA was observed in density fractions from 1-1.15 g/ml with a peak at 1.09-1.10 g/ml [5,6]. In contrast, there was no specific low density peak of HCV RNA in the media of MMH1-1flJFH#2 cells and total viral RNA levels in the supernatant were significantly lower than that seen the in Huh-7 replicon cell supernatant. Hence, these results support the conclusion that mouse hepatocytes do not efficiently secrete enveloped, flJFH-1 RNA-containing particles.

Discussion
In this study, we have shown that replicon clones derived from the HCV JFH-1 consensus cDNA can establish and maintain efficient replication in both mouse hepatocytes (e.g. MMHD3, MMH1-1, AML12) and mouse embryonic fibroblasts (e.g. NIH3T3). This discovery reveals a previous unrealized breadth in the host range of the JFH-1 HCV clone, and expands our ability to develop experimental HCV model systems. In addition to observing stable expression of viral proteins and replication of viral RNA (Fig. 2), we found that HCV replication in mouse hepatocytes is not dependent on adaptive mutations in the viral genome (Fig. 3), but is sensitive to inhibition by interferons (data not shown) and the HCV-specific small molecule inhibitor, BILN2061 (data not shown). Taken together, these results demonstrate replication of nonadapted HCV replicon clones in mouse cells, and they potentially provide the basis for the production of a mouse model of HCV replication.
While Zhu et al., (2003) previously provided evidence that HCV replication in mouse hepatocytes was possible, the data presented here differs from that report in three fundamental ways. First, Zhu et al., (2003) could only initiate HCV replication in mouse cells by transfection with a HCV RNA quasispecies obtained from previously established HeLa replicon cells, but they were not able to establish replication in mouse cells with any individual replicon clone. In contrast, we have been able to identify a defined HCV replicon clone capable of replicating in various mouse cell lines. The identification of a specific mouse-permissive HCV clone is advantageous because it facilitates mouse model development and allows for Sequence analysis of sgJFH-1 replicons in mouse hepatocytes reverse-genetics experimentation. Second, whereas replication of the subgenomic 1b HCV-N replicon reported by Zhu et al., (2003) was restricted to one specific mouse cell line, we have been able to establish JFH-1 HCV replication in multiple mouse cell lines significantly enhancing the likelihood that future in vitro and in vivo experimental HCV mouse models based on the JFH-1 clone might be possible. Finally, while the previous success was dependent on cell culture adaptive mutations that prevent HCV infectivity [23], we have demonstrated that replicons derived from the JFH-1 consensus sequence can replicate in mouse hepatocytes, in the absence of adaptive mutations that might otherwise alter the infectivity of the clone.
The ability of the JFH-1 clone to replicate efficiently in tissue culture in the absence of adaptive mutations has recently led to the development of in vitro HCV infection models in Huh-7 cells [5][6][7] and the demonstration that full length HCV replicons are capable of producing infectious viral particles (Fig. 5B-C). Likewise it provided us the opportunity to study the ability of mouse cells that support JFH-1 replication to uptake and secrete infectious HCV JFH-1 particles. Consistent with previous HCV pseudoparticles data, we observed no evidence that HCV particles were able to productively enter mouse cells, even when we used G418 treatment to select for cells that might have become HCV positive following inoculation. More unexpectedly, we also observed that mouse hepatocytes replicating infectious JFH-1 RNA were not able to secrete infectious HCV RNA containing particles. Though it is possible there were low levels of infectious particles present that were below the detection limits of our assays, sucrose density gradient analysis and RT-QPCR analysis of HCV RNA isolated from the mouse culture supernatants supports the conclusion that very few, if any, enveloped, HCV RNA-containing particles were being secreted from these mouse cells (Fig. 5C). This may reflect an inability to assemble infectious particles, secrete viral particles, or both. Hence, in addition to productive viral entry, infec-Northern Blot detection of sgJFH-1 RNA in AML12 mouse hepatocytes and NIH3T3 mouse fibroblasts in the culture medium of the same MMH and Huh-7 clones was determined by incubating serial dilutions on naïve Huh-7 cells and performing immunofluorescence for HCV 3 days post-inoculation, as previously described [5]. The titer is expressed as NS5A-positive focus-forming units (ffu) per ml of medium. (C) RT-QPCR detection of extracellular HCV RNA following sucrose density gradient centrifugation.
tious particle production appears to be another aspect of the HCV life cycle that may not be readily recapitulated in mouse hepatocytes. Clearly, elucidating the basis of this phenomenon may help us understand the key steps required for infectious HCV particle secretion in human cells.

Conclusion
The data presented in this report demonstrate that JFH-1 derived HCV replicon clones can efficiently replicate in mouse cells. While this discovery certainly represents an advance in our ability to develop mouse models of HCV replication, further analysis of infectious JFH-1 HCV entry into and egress from mouse cells also identified at least 2 other aspects of the HCV life cycle that are not supported by the mouse hepatocytes tested. Such species restrictions will likely continue to make the development of a purely murine HCV infectious model challenging.

HCV constructs
Constructs containing the non-adapted wild type (pHCVrep1b BartMan_AvaII) and Huh-7 cell adapted [pHCVrep1bBB7 (S1179I)] HCV Con1 genotype 1b subgenomic replicons [9] and the genotype 1a H77 subgenomic replicon [H/SG-Neo(L+I)] [22] were provided by Dr. Charles Rice (Rockefeller University, NY). The pNNeo/3-5B(SI) construct encoding the genotype 1b HCV-N subgenomic replicon was provided by Dr. Stanley Lemon (University of Texas Medical Branch, Galveston) [13]. The constructs pSGR-JFH1, pFGR-JFH1, and pJFH-1 containing the sgJFH-1 HCV genotype 2a JFH-1 replicon, the full length (fl)JFH-1 replicon, and the genomic JFH-1 viral clone have been described [7,24,25]. In all constructs, the HCV cDNA is located at the 1+ position 5' of the T7 promoter. Plasmids were linearized at the 3' end of the HCV cDNA and used as a template for in vitro transcription by T7 RNA polymerase (MEGAscript; Ambion, Austin, TX). To generate strand-specific RNA probes, a 1 kb fragment of the JFH-1 NS5B coding region (HindIII-EcoRV) was cloned into the pBSKII+ vector to allow for T7 and SP6-driven transcription of JFH-1 negative and positive strand probes, respectively.  [18]. This immortalized mouse hepatocyte cell line was derived from the liver of transgenic mice, which express the constitutively active cytoplasmic domain of the human hepatocyte growth factor receptor (cMet) in their livers. The MMH1-1 cells were independently derived from double transgenic cMet transgenic mice that also have a hepatitis B virus (HBV) transgene [19]. Although these cells contain a HBV transgene, expression and replication of HBV in these cells only occurs after the cells become confluent and are further differentiated in the presence of 2% DMSO for 8 days [19]. All of the experiments presented here were performed on subconfluent cell cultures in the absence of DMSO; therefore, no HBV replication was occurring in these cells (data not shown). Both MMHD3 and MMH1-1, were plated on collagen I Biocoat dishes (Becton Dickinson, Franklin Lakes, NJ) in RPMI 1640 (Gibco Invitrogen) supplemented with 10% fetal calf serum (Gibco Invitrogen), 55 ng/ml EGF (Becton Dickinson), 16 ng/ml IGF-II (Calbiochem, San Diego, CA), 10 μg/ml insulin (Sigma) and Penicillin-Streptomycin-Glutamine (100×, liquid)(Gibco Invitrogen) [19]. G418 was added to culture media as indicated (Invitrogen).

HCV RNA transfection and G418-resistant colony formation
In vitro transcribed HCV RNA or total cell RNA was transfected into cells using a modified electroporation protocol [10]. Trypsinized cells were washed twice with serum-free medium and resuspended to a final concentration of 1 × 10 7 cells/ml. One to ten micrograms of HCV RNA was then mixed with 0.4 ml of the cells in a 4 mm cuvette. A Gene Pulser system (BioRad Laboratories, Hercules, CA) was used to deliver the following single pulses: Huh-7 (0.27 kV, 100 OHMS, 960 μF); Met-based and AML12 cells (0.45 kV, 100 OHMS, 960 μF); NIH3T3 (0.27 kV, 100 OHMS, 960 μF). For the generation of replicon clones, transfected cultures were maintained in the presence of G418 at a concentration of 500 μg/ml until all cells died or distinct G418-resistant cell colonies formed.
To visualize colony formation, cells were fixed and stained with crystal violet.

RNA analysis
Total cell RNA was isolated by the guanidine thiocyanate method using standard protocols [26]. RNA was resolved in 1% agarose, 2.2 M formaldehyde gels and transferred to nylon membrane (Schleicher & Schuell, Keene, NH). Membranes were cut across the 28S ribosomal band so that the top of the blot could be hybridized with 32 Plabeled strand-specific riboprobes (MAXIscript; Ambion), while the bottom was hybridized with 32 P-labeled cellular GAPDH cDNA probe (Random Prime Synthesis; Invitro-