HepG2 Cells Expressing MicroRNA miR-122 Support the Entire Hepatitis C Virus Life Cycle

Abstract

The liver-specific microRNA miR-122 is required for efficient hepatitis C virus (HCV) RNA replication both in cell culture and in vivo. In addition, nonhepatic cells have been rendered more efficient at supporting this stage of the HCV life cycle by miR-122 expression. This study investigated how miR-122 influences HCV replication in the miR-122-deficient HepG2 cell line. Expression of this microRNA in HepG2 cells permitted efficient HCV RNA replication and infectious virion production. When a missing HCV receptor is also expressed, these cells efficiently support viral entry and thus the entire HCV life cycle.

TEXT

Hepatitis C virus (HCV), a member of the family Flaviviridae, is associated with more than half of newly diagnosed hepatocellular carcinomas in the United States (1, 31) and is the leading cause of liver transplants worldwide (7). Historically, a lack of model systems to study HCV replication has slowed the progress of HCV research and the development of specific antivirals. Few HCV genomes replicate in culture, and only one efficiently produces infectious particles (23, 39, 42). Most model systems rely on the Huh-7 cell line and derivatives, such as the highly HCV-permissive Huh-7.5 subclone (6). The development of novel cell systems capable of supporting the entire HCV life cycle would allow broader examination of interactions between HCV and host cell biology and would provide essential tools for anti-HCV drug discovery.

MicroRNAs (miRNAs) are small RNA molecules predicted to downregulate the expression of one-third of human genes by reducing mRNA stability and/or translation, depending on the degree of complementarity between the miRNA and the mRNA target site (2). miR-122, a liver-specific miRNA expressed at high levels in hepatocytes, is required to support HCV RNA replication (19). Two miR-122 target sites have been identified in the 5′ untranslated region of the HCV genome, and mutation of these sites disrupts HCV RNA replication. Replication is restored by cotransfection with RNA duplexes that produce a mutant miR-122 with complementary changes (16–19, 26). How miR-122 enhances HCV replication is unclear. Although its binding appears to enhance HCV internal ribosome entry site (IRES)-directed translation, this effect is not great enough to account for the difference in RNA replication (16, 17, 40). Furthermore, the normal cellular function of miR-122 in regulating cholesterol and fatty acid biosynthesis is not required for this enhancement (18). Thus, miR-122 has been hypothesized to directly affect HCV RNA replication or stability (17, 19, 26). Although inhibiting miR-122 has been shown to be an effective therapy in chimpanzees infected with HCV (21), further development of such therapies will require a firmer understanding of how miR-122 promotes HCV replication.

The majority of published studies on the role of miR-122 in the HCV life cycle have described research conducted with HCV-permissive Huh-7 cells or derivatives, which express endogenous miR-122 (19). Two studies explored the effect of miR-122 on HCV in divergent cell types and found that miR-122 enhances intracellular HCV RNA replication 6- to 11-fold in kidney-derived HEK293 cells (8) and 40-fold in mouse embryonic fibroblasts (22). Neither of these studies tested these systems' capacities to support the entire HCV life cycle, including early and late events such as cell entry and infectious virion release. We sought to examine how miR-122 affects the HCV life cycle in the hepatocellular carcinoma-derived HepG2 cell line. This cell line does not express endogenous miR-122 (19) and only weakly supports HCV replication (at a level ∼850-fold lower than that in Huh-7.5 cells), which is insufficient for efficient viral assembly (12, 23). Establishing efficient HCV replication in HepG2 cells would be advantageous as these cells, in contrast to Huh-7 cells, polarize (reviewed in reference 10) and would thus permit the examination of how cell polarity impacts the HCV life cycle.

miR-122 expression enhances HCV RNA replication in HepG2 cells.

To express miR-122 in HepG2 cells, we transfected these cells with a plasmid carrying the neomycin resistance gene and the green fluorescent protein (GFP) gene with a beta-globin intron insertion containing the human genomic miR-122 precursor sequence (Fig. 1A) (the sequence is available upon request). As miRNAs are usually expressed within noncoding or intronic RNA transcribed by RNA polymerase II (3), this context ensures proper processing of mature miR-122. Approximately 5.4 × 10⁶ HepG2 cells seeded on collagen were transfected with 3 μg of this plasmid using TransIT-LT1 transfection reagent (Mirus). Cells were selected in medium (Dulbecco's modified Eagle medium [DMEM] with 10% fetal bovine serum [FBS]) containing 400 μg/ml G418 (Invitrogen) for several weeks to generate over 1,000 individual colonies, which were pooled and dilution cloned in 96-well plates. Total RNA was prepared from expanded single-cell, GFP-positive clones by extraction with TRIzol (Invitrogen). Northern blot analysis using previously described methods (32, 33), with equal quantities of RNA and U6 RNA as a loading control, showed that all clones expressed detectable levels of miR-122 and did not exhibit defects in general miRNA processing as determined by blotting for the unrelated miRNA miR-93 (Fig. 1B).

Fig. 1.

miR-122 enhances HCV RNA replication in HepG2 cells. (A) Illustration of the plasmid used to stably express miR-122 in HepG2 cells. Two expression cassettes are shown. The first carries the GFP gene bearing an intron, flanked by splice donor (SD) and splice acceptor (SA) sites, containing the miR-122 genomic locus and precursor sequence. The second carries the neomycin resistance gene (NeoR). HCMV, human cytomegalovirus; SV40, simian virus 40. (B) Northern blots for miR-122, miR-93, and U6. Nucleotide sizes (in bases) are indicated to the left of each blot. (C) Illustration of HCV JFH-1 subgenomic replicon organization. The positions of the GLuc reporter gene, the EMCV IRES, and the HCV nonstructural (NS) protein genes are marked. The locations of the NS5B polymerase active site (GDD) and inactivating mutations (GNN) are also indicated. (D) Quantification of GLuc relative light units (RLU) produced at the indicated hours posttransfection (hpt) by transfection of Huh-7.5 cells, as a positive control, and naïve and miR-122-expressing HepG2 cell clones with either a replication-defective (GNN) or the wild-type (WT) subgenomic HCV replicon. In this and all subsequent luciferase assays, the background value from naïve cell supernatants was 55 ± 17 RLU (mean ± standard deviation [SD]). (E) The indicated cell populations were cotransfected with the subgenomic HCV replicon illustrated in panel C and either no RNA (mock), a random-sequence oligonucleotide, or a miR-122-complementary 2′-O-methylated RNA oligonucleotide (miR-122). GLuc activity was assayed at the indicated time points. (F) Diagram of an miRNA lentiviral expression construct with an internal cytomegalovirus (CMV) promoter to express transcripts carrying either the blasticidin (BsdR) or puromycin (PuroR) resistance gene followed by the miR-122 genomic locus. LTR, long terminal repeat; Psi, HIV-1 packaging signal; RRE, Rev response element; cPPT, central polypurine tract; LTRdelU3, long terminal repeat with a deletion in the U3 region. (G) Naïve HepG2 cells were transduced with lentiviral pseudoparticles bearing the above-mentioned provirus to express either wild-type or mutant (3-4 MT) miR-122. Following drug selection, these cell populations were transfected, in parallel with naïve cells, with GLuc-encoding subgenomic HCV replicon RNA and cells were cultured in either dimethyl sulfoxide (DMSO) or a 6 μM concentration, which is 50 times the 50% inhibitory concentration (IC₅₀), of the HCV polymerase inhibitor 2′-C-methyladenosine (2′CMA), to substitute for a GNN control. GLuc activity was assayed at the indicated time points. Results are expressed as means ± SD (n = 3 independent transfections) and are representative of results from multiple experiments.

The capacity of these cells to support HCV RNA replication was tested by transfection with 2 μg of in vitro-transcribed replication-competent (wild-type) and polymerase-defective (GNN) HCV subgenomic replicon RNAs derived from the HCV genotype 2a JFH-1 genome (Fig. 1C) (20) as described previously (5). In these replicons, the HCV IRES directs translation of the Gaussia princeps luciferase protein (GLuc) (37, 38), which was quantified using the Renilla luciferase assay system (Promega) as described previously (30). Although all transfections produced detectable GLuc, the wild-type replicon signal amplified over time, which is indicative of RNA replication (Fig. 1D). Even in naïve HepG2 cells, RNA replication was detectable and statistically significant (compare data for transfections of naïve HepG2 cells with GNN and wild-type replicons at 48 h posttransfection [P = 0.0059]). At 48 h posttransfection, when replication appears to plateau, miR-122-expressing HepG2 clones supported 50- to 125-fold greater GLuc activity than naïve HepG2 cells (P < 0.0075). Thus, miR-122 expression enhances HCV subgenomic replicon replication in HepG2 cells. miR-122 levels do not correlate with replication efficiency, suggesting that even low quantities of miR-122 can greatly enhance HCV replication. Notably, the level of GLuc expression from nonreplicating GNN RNAs was higher at all time points in HepG2 cells expressing miR-122 (referred to herein as HepG2 miR-122 cells) than in HepG2 parental cells, suggesting that this miRNA enhanced translation from the HCV IRES, HCV RNA stability, or both. The enhancement of HCV RNA replication in HepG2 miR-122 cells was due to expression of this miRNA, as cotransfection with a miR-122-specific 2′-O-methylated RNA oligonucleotide antagomir, but not a random-sequence antagomir (19), resulted in a significant reduction in replicon reporter expression (Fig. 1E). This enhancement also required both wild-type miR-122 and complementary target sites within the HCV genome, as the replication of an HCV replicon with mutant 5′-untranslated-region miR-122 seed sequences, termed p3-4 in a prior publication (19), was not enhanced by wild-type miR-122 and expression of the complementary mutant p3-4 miR-122 sequence did not enhance replication of the wild-type replicon (Fig. 1G). Although expression of the p3-4 mutant miRNA did permit replication of the mutant replicon, it was not as efficient as replication with the wild-type pairing.

HCV replication in HepG2 cells is limited by transfection efficiency.

While miR-122 expression in HepG2 cells enhanced HCV RNA replication, this process was still on average 11-fold less efficient in these cells than in Huh-7.5 cells (Fig. 1D). To examine the frequency of sustained HCV replication, transfected cells were trypsinized and immunostained for HCV NS5A by using the 9E10 mouse monoclonal antibody (23) and a goat anti-mouse Alexa Fluor 647 secondary antibody (Invitrogen). At 72 h posttransfection, 32% of transfected Huh-7.5 cells were NS5A positive, as determined by fluorescence-activated cell sorter (FACS) analysis, while naïve and miR-122-expressing HepG2 cell populations exhibited 1.2 and 6.3% NS5A-positive cells, respectively (Fig. 2A). As the HCV proteins in this context are expressed from the encephalomyocarditis virus (EMCV) IRES rather than the HCV IRES, these results suggest that miR-122 enhances HCV RNA replication and not NS5A translation.

Fig. 2.

Transfection efficiency limits HCV replication in HepG2 cells. (A) Example and quantification from three independent transfections (results shown are means ± SD) by FACS analysis of intracellular NS5A staining within the indicated cell populations 72 h posttransfection with HCV replicons. (B to F) Huh-7.5 cells, as a positive control, and naïve and miR-122-expressing HepG2 cells were cotransfected with GLuc-expressing wild-type and polymerase-defective (GNN) HCV subgenomic replicons and capped FLuc RNA. (B and C) Assays for secreted GLuc (B) and intracellular FLuc (C) were conducted at the indicated time points posttransfection. Results shown are means ± SD (n = 3 independent transfections) and are representative of results from two independent experiments. (D) FLuc values were used to determine transfection efficiencies for the indicated cell populations, calculated as the average ratio of FLuc activities between cell populations at each time point and presented relative to values for Huh-7.5 cells. (E) Graph of GLuc values normalized for relative transfection efficiency to allow comparison of HCV replication efficiencies independently of transfection efficiency. Symbols are as defined in Fig. 3B. (F) Evaluation of HCV IRES activity in the various cell populations, presented as the ratios of GLuc to FLuc measured from each GNN transfection at the indicated time points and normalized to values determined for naïve HepG2 cells. CAP, Cap-dependent translation.

To more accurately compare the relative capacities of these cells to support HCV replication, cells were cotransfected with HCV subgenomic replicon RNA and, as a transfection control, an in vitro-transcribed, capped, and polyadenylated mRNA encoding the firefly luciferase protein (FLuc) constructed identically to one used in a previous study (17). HCV IRES-based GLuc expression results were similar to those described above (Fig. 2B). FLuc activity, quantified using the luciferase assay system (Promega) as described previously (11), was detected in all samples, with peak expression between 4 and 6 h posttransfection (Fig. 2C). At each time point, naïve and miR-122-expressing HepG2 cells had similar FLuc expression levels and thus have comparable transfection efficiencies, which we calculated as the average ratio of FLuc activities at each time point (Fig. 2D). Huh-7.5 cells were more easily transfected than any of the HepG2 cell populations, as FLuc levels in Huh-7.5 cells were 6.5- to 9-fold higher than those in HepG2 cells (Fig. 2D). After normalizing the GLuc values at each time point for transfection efficiency, there was little difference between the abilities of Huh-7.5 and miR-122-expressing HepG2 cells to support HCV RNA replication (Fig. 2E).

To directly compare the efficiencies of HCV IRES-dependent translation in the above-mentioned cell types, the GLuc values from GNN replicon transfections were normalized to the respective time point's FLuc values. Huh-7.5 and naïve HepG2 cells exhibited similar abilities to support HCV IRES-dependent translation (Fig. 2F). Furthermore, HCV IRES-dependent translation was only 1.4- to 2.1-fold more efficient in miR-122-expressing HepG2 cell clones than in the parental HepG2 cell population. Together, these results support the hypothesis that translation is not greatly impacted by miR-122 expression in HepG2 cells (17).

HepG2 miR-122 cells can be engineered to support HCV cell entry.

With the use of lentiviral pseudoparticles bearing the HCV envelope glycoproteins (HCVpp), HepG2 cells have been shown previously to support HCV cell entry if the missing receptor, CD81, is overexpressed (4, 12, 41). To examine HCV cell entry in miR-122-expressing HepG2 cells, naïve and miR-122-expressing HepG2 cells were transduced with a lentivirus to stably express the human CD81 protein, according to previously published protocols (30). Transduction efficiency and CD81 cell surface expression were evaluated by staining of nonpermeabilized cells with an antibody reactive to an extracellular CD81 epitope followed by FACS analysis (Fig. 3A) conducted as described previously (12). To determine their capacity to support HCV cell entry, these cells were challenged with HCVpp and, as a positive control, pseudoparticles bearing the vesicular stomatitis virus envelope glycoprotein (VSVGpp). As shown previously, naïve HepG2 cells are not susceptible to HCVpp infection, in contrast to Huh-7.5 cells, and this deficiency is eliminated by CD81 expression (Fig. 3B). All miR-122-expressing HepG2 clones were similarly susceptible to infection with HCVpp when CD81 was expressed.

Fig. 3.

Testing of HCV cell entry capacities. (A) FACS analysis of CD81 expression on the surfaces of the indicated cells. For each cell type, data for isotype control (gray filled) and CD81 JS-81 antibody (unfilled) staining are shown. The y axis represents the percent maximum. (B) Luciferase production from the cell populations indicated in panel A, as well as Huh-7.5 cells as a positive control, 48 h postinfection (hpi) with lentiviral particles bearing the HCV (white bars) and vesicular stomatitis virus (black bars) envelope proteins (means ± SD of readings from 4 independent infections). (C) Illustration of the Jc1 chimeric HCV genome. The region comprising the core through the end of the NS2 genes corresponding to the amino terminus (light gray) is from the HC-J6 genome, and the remaining NS genes (dark gray) are from the JFH-1 HCV isolate (34). (D) The effective titer of a single preparation of Jc1 HCVcc was determined on the indicated cell populations. The dotted line indicates the level of detection of the assay. Results shown are means ± SD of readings from 3 independent infections. TCID₅₀, 50% tissue culture infective dose.

To gauge the capacity to support HCVcc infection, the effective titer of a single stock of HCVcc for each cell population was determined by limiting dilution assay, the results of which were quantified by NS5A staining as described previously (23). As shown in Fig. 3D, infection of naïve HepG2 cells, expressing neither CD81 nor miR-122, was below the level of detection of this assay. HepG2 cells expressing CD81 but not miR-122 were 467-fold less infectible with HCVcc than Huh-7.5 cells. The susceptibility to HCVcc infection was increased another 22- to 77-fold by miR-122 expression in HepG2 cells transduced with CD81, to within 6- to 20-fold that of naïve Huh-7.5 cells.

miR-122 expression in HepG2 cells permits efficient infectious HCV release.

To test the ability to support infectious HCV assembly and release, we transfected Huh-7.5, HepG2 CD81, and HepG2 CD81/miR-122-expressing cells with full-length bicistronic HCV RNAs that express the GLuc protein from the HCV IRES (Fig. 4A). Similar to the results of the above-described subgenomic replicon experiments, miR-122 expression enhanced HCV replication in HepG2 cells. These HCV RNAs exhibited more robust replication in the miR-122-expressing HepG2 cells than in naïve HepG2 cells, as the GLuc levels from miR-122-expressing cells were consistently 10- to 36-fold higher (Fig. 4B). Supernatants from these cultures were collected, filtered, and used to infect Huh-7.5 cells to determine the relative amounts of infectious HCVcc released. All cell populations produced infectious virus, and cell populations expressing miR-122 yielded 30- to 71-fold more infectious virus than HepG2 cells not expressing this miRNA (Fig. 4C). Thus, miR-122 expression enhanced the complete HCV life cycle in HepG2 cells. To examine the kinetics of both viral spread and virion release, Huh-7.5 and HepG2 CD81/miR-122 cells were infected with nonreporter HCVcc at a low multiplicity of infection (MOI) and the percentage of NS5A-positive cells and the amount of infectious virus release were measured over time. Although virus spread more efficiently within Huh-7.5 cell cultures, the percentage of NS5A-positive cells did increase in HepG2 CD81/miR-122 cell cultures, and both populations released infectious virus. In this experiment, HCV infection induced overwhelming Huh-7.5 cell death beginning at 5 days postinfection, which makes it difficult to directly compare the amounts of HCVcc produced from the different cell populations. It is unclear why HCVcc did not spread more efficiently within the HepG2 CD81/miR-122 culture. As the results of the above-described experiments suggest that RNA replication is equally efficient in HepG2 CD81/miR-122 cells and Huh-7.5 cells, it is likely that deficiencies in cell entry, perhaps due to cell polarity, or the presence of a more intact interferon response pathway limits infection within HepG2 cells.

Fig. 4.

miR-122 enhances replication and infectious virus release in HepG2 cells. (A) Illustration of the bicistronic GLuc-expressing Jc1 chimeric HCV genome. (B) The indicated cells were transfected with full-length HCVcc reporter genomic RNA. To measure RNA replication, reporter activity produced by transfected cells was determined at the indicated time points posttransfection. (C) Filtered supernatants from these transfections were applied to Huh-7.5 cells to determine relative quantities of infectious virus produced by each transfection at the indicated time points. These infections were assayed 48 hpi. (D) The indicated cells were infected with full-length, nonreporter Jc1 HCVcc RNA in parallel. At the indicated time points, supernatants were collected and the quantity of infectious virus release was determined by a TCID₅₀ assay on Huh-7.5 cells (gray lines; right y axis). At the same time points, portions of the cell populations were passed into fresh plates and analyzed for the presence of HCV by NS5A FACS analysis (dark lines; left y axis). Results shown are means ± SD of readings from 3 parallel infections.

In summary, we have demonstrated that HepG2 cells support efficient HCV RNA replication and infectious virion release if miR-122 is expressed. The further addition of the HCV entry factor CD81 renders these cells able to support the entire viral life cycle. Our results support the notion that miR-122 is required for HCV RNA replication but does not greatly enhance viral translation. Although a mechanism for this enhanced RNA replication is still not clear, it may occur through stabilization of the HCV genome by shielding of the viral 5-triphosphate from exonucleases or retinoic acid inducible gene I (RIG-I) surveillance, as recently hypothesized by Machlin et al. (26). Our study does not provide evidence that miR-122 enhances any other stage of the viral life cycle. Although several non-Huh-7-based HCV replication systems have been developed, only Huh-6 cells have been shown to be capable of secreting infectious HCV (14). HepG2 cells offer a distinct advantage over these cell lines, as they can grow in a polarized manner that mimics the bile canalicular configuration of hepatocytes in vivo and thus provide a model system that more closely resembles the natural host cell for HCV (reviewed in reference 10). Two essential HCV cell entry factors, claudin-1 (11) and occludin (25, 35), are tight-junction proteins and are not likely to exhibit completely conserved functions in nonpolarized cells. Thus, a polarized-cell model is particularly important for the study of the HCV cell entry process. HepG2 cells have been used in studies of HCV polarized-cell entry using HCVpp (15, 27, 28); however, the scope of such studies is limited by the low titer of these particles and the likelihood that their entry pathways do not perfectly follow those of native HCV. The ability to infect HepG2 cells with HCVcc, which can be efficiently concentrated and purified (13, 24, 29, 36) and is amenable to fluorescent labeling (9), opens the way to conducting detailed analysis of HCV cell entry pathways in polarized cells.