|Home | About | Journals | Submit | Contact Us | Français|
The recent development of a cell culture model of hepatitis C virus (HCV) infection based on the JFH-1 molecular clone has enabled discovery of new antiviral agents. Using a cell-based colorimetric screening assay to interrogate a 1,200-compound chemical library for anti-HCV activity, we identified a family of 1,2-diamines derived from trans-stilbene oxide that prevent HCV infection at nontoxic, low micromolar concentrations in cell culture. Structure-activity relationship analysis of ~300 derivatives synthesized using click chemistry yielded compounds with greatly enhanced low nanomolar potency and a >1,000:1 therapeutic ratio. Using surrogate models of HCV infection, we showed that the compounds selectively block the initiation of replication of incoming HCV RNA but have no impact on viral entry, primary translation, or ongoing HCV RNA replication, nor do they suppress persistent HCV infection. Selection of an escape variant revealed that NS5A is directly or indirectly targeted by this compound. In summary, we have identified a family of HCV inhibitors that target a critical step in the establishment of HCV infection in which NS5A translated de novo from an incoming genomic HCV RNA template is required to initiate the replication of this important human pathogen.
Hepatitis C virus (HCV) is a hepatotropic, enveloped virus with a single-stranded RNA genome of positive polarity that causes acute and chronic hepatitis and hepatocellular carcinoma (36). More than 170 million people are chronically infected and are at risk of developing cirrhosis and hepatocellular carcinoma (25). There is currently no vaccine against HCV, and the only approved antiviral therapy, a combination of recombinant type I alpha interferon and ribavirin, is expensive, toxic, and only partially effective (50).
HCV is a member of the Flaviviridae family. Its 9.6-kb RNA genome carries a long open reading frame that is co- and posttranslationally cleaved by cellular and viral proteases (reviewed in reference 52) into structural proteins (core, E1, and E2) that constitute the major viral components of the viral particles and nonstructural proteins (p7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B) that are required at multiple levels of the virus life cycle, including viral RNA replication (2) and infectious-particle assembly (46). The single open reading frame is located between two untranslated regions (UTRs), the 5′ UTR and the 3′ UTR, which contain RNA sequences essential for RNA translation and replication, respectively (15, 16, 24).
HCV infection is initiated when viral particles bind and enter the target cells in a process that involves multiple cellular receptors and clathrin-dependent endocytosis (7). Release of the viral genome into the cytoplasm is thought to occur after low-pH-induced membrane fusion, a process mediated by the viral envelope glycoproteins (60). Incoming viral genomes are translated into the viral polyprotein in a process that requires components of the autophagy machinery (12). Cleavage of the polyprotein into the individual viral proteins enables the establishment of replication complexes in endoplasmic reticulum (ER)-derived membranous compartments where viral RNA replication occurs via minus-strand synthesis (45). Progeny genomes are either translated to produce additional viral proteins or packaged to assemble progeny infectious virions. Virus particle assembly is thought to occur in an ER-related compartment in close proximity to cytosolic lipid droplets, where core and NS5A proteins colocalize (41, 44). HCV RNA-containing core particles acquire their envelope by budding through the ER membrane at sites where E1/E2 glycoprotein complexes are inserted (54). The HCV assembly process is dependent not only on structural and nonstructural proteins (p7, NS2, NS3, and NS5A) (1, 27, 29, 30, 39, 43, 47, 55, 58, 62) but also on cellular factors involved in lipoprotein biosynthesis (10, 18, 26, 28, 48). HCV assembly results in the production of high-density intracellular infectious precursors that undergo maturation during their passage through the secretory pathway, where they acquire their characteristic low-density extracellular configuration (21). Extracellular virions are spherical, pleomorphic particles that are heterogeneous in size (31, 35, 42, 53, 59), some of which are surrounded by a membrane bilayer, likely the viral envelope, that can be observed by cryoelectron microscopy (19).
The development of hepatoma-derived cell lines bearing autonomously replicating HCV RNA (replicons) (8, 38) permitted the development of potent antiviral molecules directed against the viral NS3 protease, the NS5B RNA-dependent RNA polymerase (49), and recently NS5A (17), a zinc metalloprotein that is required for viral replication and particle assembly (41, 58). The opportunity to develop compounds that target additional steps in the HCV life cycle awaited the development of cell culture infection models that recapitulated all aspects of the viral infection. We recently established a cell-based, unbiased screening procedure that permits interrogation of chemical libraries for antiviral activity against all steps in the HCV life cycle, including cellular and viral factors required for efficient viral spread (22). Using this technology, we have discovered a class of antiviral compounds that selectively blocks the initiation of HCV RNA replication in vitro without affecting ongoing HCV RNA replication, a unique and hitherto obscure step in HCV infection that can now be approached experimentally for basic and translational purposes.
The initial chemical library was composed of ~1,200 compounds with diverse skeletal architectures (i.e., acyclic, cyclic, fused/bridged polycyclic, heterocyclic, and aromatic). Their assembly relied on click chemistry, focusing primarily on carbon-heteroatom linkages (32). This synthetic strategy allowed for rapid generation of large quantities (grams) of material in high yields requiring relatively simple purification methods. Inherently incorporated into these structures were functional groups that immediately enabled further diversification for structure-activity relationship (SAR) analysis and hit compound optimization.
The syntheses of compounds 1 and 2 are illustrated elsewhere (http://www.scripps.edu/wieland/data/Gastaminza_2/Suppl.FIg1.pdf). Synthetic procedures for compound 1 have been described previously (20). A detailed preparation of compound 2 along with the SAR studies leading to compound 2 will be described in a future report (unpublished data). Dry powders of compounds 1 and 2 were dissolved in dimethyl sulfoxide (DMSO) to a final 10 mM concentration. The solution of compound 2 required heating at 65°C for 5 min before use, as this compound tends to precipitate in DMSO at a relatively high concentration (10 mM), although it is water soluble up to 50 μM.
The chemical library described above was screened using a colorimetric HCV infection readout as described previously (22). Briefly, compounds were dissolved in DMSO at a final concentration of 10 mM, diluted to a final concentration of 20 μM in 100 μl of culture medium (Dulbecco's modified Eagle's medium containing 10% fetal calf serum [DMEM-10% FCS]), and mixed (1:1) with 100 μl of a D183 virus (65) dilution (2 × 103 focus-forming units [FFU]/ml) in medium. One hundred microliters of this mixture was used to inoculate 104 Huh-7.5.1-clone 2 (Huh-7.5.1-c2) cells (51) per well (multiplicity of infection [MOI] of 0.01) in a 96-well format (Falcon flat-bottom 96-well cell culture microplate; BD Biosciences, La Jolla, CA) in duplicate, as previously described (20). The cells were incubated at 37°C for 72 h, after which the cells were fixed with 4% paraformaldehyde (PFA) for 20 min at room temperature. PFA-fixed cells were subsequently processed as described previously (22). Every test plate included control wells with uninfected cells that were used to subtract background values. Relative infection efficiency values were calculated from the colorimetric values using a standard curve generated by serial 2-fold virus dilutions starting at 200 FFU/well. Data were considered only if the standard curves displayed correlation coefficients (r2) above 0.97.
Compound toxicity was determined by evaluating the cell biomass remaining at 72 h postinoculation by crystal violet staining and colorimetry at 570 nm as described previously (6). Compounds resulting in a reduction of the biomass below ~70% of that of the controls were considered toxic and discarded for further analysis. Compounds reducing infection efficiency more than 10-fold in the absence of toxicity were considered hits and were further characterized.
Compound stock solutions (100 μM) were prepared by diluting a 10 mM solution in DMSO (1.5 μl) in complete medium (148 μl). Subsequent serial 3-fold dilutions of the compound were prepared in culture medium (DMEM-10% FCS). Compound dilutions (50 μl) were mixed with 50 μl of a virus dilution in complete medium containing 200 FFU mixtures containing virus, and compound dilutions ranging from 50 μM to 0.4 nM were added to Huh-7.5.1-c2 cells in 96-well plates and assayed as described above. Values of 50% effective concentration (EC50) and EC90 were obtained by graphic interpolation of the compound concentration resulting in 50% and 90% HCV infection inhibition in the linear portion of the curve. Values of 50% lethal dose (LD50) were determined by evaluating the cell biomass remaining at 72 h postinoculation as described above. LD50 values were interpolated graphically from the dose-response curves. Similar values were obtained using MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide]–formazan cellular viability assays performed in parallel (23).
Huh-7, Huh-7.5.1, and Huh-7.5.1-c2 cells were cultured in DMEM-10% FCS as previously described (51, 64). JFH-1 and D183 adapted virus stocks were produced as described previously (65). Establishment of persistent JFH-1 infections was previously described (18).
HCV E1/E2-pseudotyped lentiviral particles (HCVpp) (or, as a control, vesicular stomatitis virus G-pseudotyped lentiviral particles [VSVpp]) bearing the luciferase (luc) reporter gene were generated as described previously (4). Comparable amounts of infectious HCVpp and VSVpp were used in every experiment. For the inhibition assays, particles were mixed 1:1 with medium containing DMSO or with compound dilutions to achieve a final 5 μM concentration. The mixture was used to inoculate Huh-7 cells at 37°C for another 72-h period, after which infection efficiency was evaluated by measuring reporter gene expression using a commercial kit (luciferase assay system; Promega, Madison, WI). Relative infection values were obtained using DMSO-treated cells as a control (100%). Background levels were established by measuring luciferase expression in cells transduced with HCVpp lacking envelope glycoproteins, as previously described (4).
Huh-7.5.1-c2 cells (5 × 105 cells, 6-well plate) were inoculated (MOI of 0.01) with D183 virus (65) in the presence of 5 μM doses and incubated for 6 days at 37°C. Cells were split (1:3) at days 2 and 4 postinoculation, at which times the cells were replenished with fresh medium containing the compounds. Samples of the cells and cell supernatants were tested for infectivity and HCV RNA at days 2, 4, and 6.
Huh-7 cells were seeded in 24-well plates (5 × 104 cells/well), inoculated with D183 virus 24 h later (MOI of 10) in the presence of 5 μM doses, and further incubated for 40 h in complete medium at 37°C. Samples of the cells and supernatants were collected at 40 h postinfection for HCV RNA and infectivity analyses.
Dilutions of compound 2 (10 μM) and D183 virus (106 FFU/ml) were prepared in DMEM-10% FCS. Huh-7 cells (2.5 × 104 cells/well, 48-well plate) were either inoculated at time zero with a 1:1 mixture of the compound (250 μl) and virus dilutions (250 μl) or pretreated with the compound (1:1 mixture with medium) or the virus (250 μl of virus). At the times indicated in Fig. 2C, pretreated cells were inoculated with the virus dilution (250 μl), and cells inoculated at time zero were treated with 250 μl of the compound dilution. Cells were further incubated at 37°C for 40 h and harvested to measure HCV RNA accumulation. The compounds remained present in the medium throughout the experiment, except as shown in Fig. 2 (Wash), where the pretreated cells were washed twice with phosphate-buffered saline (PBS) before virus inoculation. Infection efficiency was determined by measuring intracellular HCV RNA accumulation 40 h postinfection.
Single-cycle infections were performed as described above by incubating virus dilutions (MOI of 10) with target cells in the presence or absence of the compounds (5 μM) for 1 h on ice. Cells were then washed twice with cold PBS. Warm medium containing either the compounds (5 μM) or DMSO as a control was added to the cells, which were further incubated at 37°C. Infection efficiency was determined by measuring intracellular HCV RNA accumulation 40 h postinfection.
A genotype 2a (JFH-1) (12) or genotype 1b (Con1-Bart79ILuc [Con1]) (11) subgenomic replicon bearing a luciferase reporter gene was electroporated into Huh-7 or Huh-7.5.1 cells, respectively, as described previously (12). For primary-translation experiments, a replication-deficient mutant JFH-1 subgenomic replicon bearing a point mutation in the NS5B catalytic site (GND mutant) was transfected (12). Cells were then plated into replicate wells (5 × 104 cells/well) in the presence of the inhibitors (5 μM) or DMSO as a negative control. Unless otherwise stated, compounds were added immediately after transfection and replenished at 24 h. Samples were collected at 5, 10, 24, and 48 h postelectroporation and tested for luciferase activity using a commercially available kit (Renilla luciferase assay system; Promega, Madison, WI).
Persistent infections were established as described previously (18). Persistently infected cells were seeded (105 cells/well, 12-well plate) the day before treatment. Compounds were diluted in medium (5 μM) and added to the cell cultures. Cells were incubated for 24 h at 37°C, after which the medium was replenished with fresh compound. After 40 h of incubation at 37°C, samples of the cells and the supernatants were analyzed for infectivity and HCV RNA.
A stock of D183 virus was used to inoculate Huh-7.5.1-c2 cells (1.5 × 106 cells, T75 flask) at a high MOI (MOI of 10) in the presence or absence of compound 2 (0.5 μM). Seventy-two hours postinfection, supernatants were collected and filtered through a 45-μm filter. Supernatants were diluted (1:10) in fresh medium and used to inoculate naïve Huh-7.5.1-c2 cells. After five blind passages in the presence of compound 2 (0.5 μM), supernatants were diluted 1:100, used to inoculate naïve cells at a low MOI in the presence of 2.5 μM compound 2, and subjected to five additional blind passages in the presence of compound 2 (2.5 μM). Diluted supernatants were used to inoculate naïve cells in the presence of 5 μM compound 2 for 10 additional blind passages in the presence of 5 μM compound. Samples of the supernatants were tested for their susceptibility to compound 2 by determining the infectivity titer in the presence or absence of 5 μM compound.
HCV RNA was isolated from the supernatants of passages 12, 14, and 18. Viral genome populations were bulk sequenced as previously described (65). All sequences were compared to the parental JFH-1 genotype, and mutations were considered only if they were present in all three samples.
Point mutants were generated by PCR-mediated mutagenesis using the following mutagenic primers (5′–3′): NsiI-RsrII fragment forward, CGCCACATGCATGCAAGCTGACCT; NsiI-RsrII fragment reverse (mutagenic E2371A), GCACGGACCGCCGGATgCGGCGG (mutations are indicated in bold, lowercase type); F2004L forward, CTGACCTCTAAATTGcTCCCCAAGCTGC; F2004L reverse, GCAGCTTGGGGAgCAATTTAGAGGTCAG; RsrII-EcoRV fragment forward, GAATCCGGCGGTCCGACGTCCCC; RsrII-EcoRV fragment reverse, GAAGAGATATCGGCCGCAAACGGCCG; K2493R forward, TCACAGAGGGCTAAAAgGGTAACTTTTGACAGG; K2493R reverse, CCTGTCAAAAGTTACCcTTTTAGCCCTCTGTGA; A2780T forward, GAGCGGAACCTGAGAaCCTTCACGGAGGCC; and A2780T reverse, GGCCTCCGTGAAGGtTCTCAGGTTCCGCTC. All PCR products were cloned into the pGEM-T vector (Promega, Madison, WI), and individual clones were sequenced to verify the presence of the mutation of interest and the absence of additional mutations. Mutant fragments were cloned into the parental vector using NsiI-RsrII sites for the F2004L and E2371A mutants and RsrII-EcoRV for the K2493R and A2780T mutants. The F2004L mutant was constructed by transferring the NsiI and BsshII fragments from the double mutant NsiI-RsrII fragment bearing the F2004L E2371A mutations into the parental plasmid. The K2493R A2780T double mutant was generated by site-directed mutagenesis of the replicon bearing the mutation A2780T.
We interrogated a small library of structurally diverse compounds by using an unbiased colorimetric cell-based screening assay for inhibitors of HCV infection that allows quantitation of JFH-1 spread in a 96-well format (20). Nontoxic compounds that strongly inhibited viral spread (>10-fold) were selected for further characterization. Analysis of the chemical structures of the antiviral compounds identified in the first round of screening led to the identification of a novel family of small molecules that significantly reduced viral spread at low concentrations (EC50 of <5 μM) in cell culture in the absence of measurable cytotoxicity (50% cytotoxic concentration [CC50] of >25 μM). The chemical structure of a representative member of the family (compound 1) is shown in Fig. 1A. Synthesis of over 300 compound 1 derivatives permitted extensive structure-activity relationship (SAR) analysis, resulting in the development of compound 2 (Fig. 1A), which displayed remarkably higher potency (EC50 of <50 nM) and comparable cytotoxicity (CC50 of >35 μM). A detailed description of the chemical library used for the screening, the SAR analysis conducted to obtain compound 2, and the structural elements required for optimal antiviral activity will be described elsewhere (unpublished data). In this report, we describe experiments that define the step of the viral life cycle that is targeted by these inhibitors and that provide insight into their mode of action against HCV in cell culture.
First, to independently confirm the ability of these compounds to interfere with HCV spread in cell culture, Huh-7 cells were inoculated at a low MOI (MOI of 0.01) with the cell culture-adapted variant of JFH-1, D183 virus (65), in the presence of compound 1 or 2 (5 μM). Extracellular-infectivity titers of supernatants collected at days 2, 4, and 6 postinfection indicated partial inhibition of HCV spread with compound 1, which caused a modest reduction in the titers found at every time point (Fig. 1B). As expected, compound 2 reduced viral titers to levels below detection (Fig. 1B), indicating that it efficiently prevented the initiation and spread of HCV infection. The antiviral activity of these compounds is underscored by the reduced intracellular HCV RNA content of compound 1-treated cultures and the complete abolition of HCV RNA replication in cells that had been pretreated with compound 2 (Fig. 1B). These results confirm the antiviral activity of these small molecules and the enhanced potency of the derivative compound 2.
Next, we performed single-cycle-infection experiments to determine the step in the viral life cycle that was targeted by the compounds. Huh-7 cells were inoculated at a high MOI (MOI of 10) in the presence of compounds 1 and 2 (5 μM). Intracellular HCV RNA levels were reduced 3-fold and >30-fold in compound 1- and compound 2-treated cells, respectively (Fig. 2A). This reduction resulted in the proportional decrease of intracellular- and extracellular-infectivity titers (Fig. 2A), indicating that compounds 1 and 2 interfere with an early step of the infection leading to the production of intracellular HCV RNA.
To examine the events preceding HCV RNA production in a single-cycle infection, we performed infections at 4°C to determine if the compounds were interfering with infection at the adsorption or postadsorption level. Huh-7 cells were inoculated at 4°C at a high MOI (MOI of 10) in the presence or absence of the antiviral compounds to permit particle adsorption without internalization, which is strongly inhibited at this temperature (33). After 1 h of incubation, the cells were washed extensively to remove free virus and compound, and infection was allowed to proceed by the addition of warm medium containing either the inhibitors or the vehicle control and by further incubation at 37°C. Analysis of intracellular HCV RNA at 40 h postinfection revealed that the compounds did not display any antiviral activity when added during particle adsorption (Fig. 2B) and that addition of the compounds postadsorption recapitulated the antiviral activity observed when the compounds were present throughout the experiment (Fig. 2B). These results suggest that the compounds target a postadsorption step in the infection and that they do not target the viral particles themselves.
In order to determine the optimal time of addition of these compounds, compound 2 was added to Huh-7 cells at different times relative to the time of inoculation (MOI of 10). Cells were pretreated (−6, −4, −2, or −1 h) with compound, or the compound was added at the time of inoculation (0 h) or after inoculation (1, 2, 4, or 6 h) and remained present throughout the experiment. Forty hours postinfection, intracellular HCV RNA levels were measured by RT-qPCR. Figure 2C shows the relative HCV RNA levels in cells treated at different times with compound 2. The curve indicates that the compound displays maximum efficacy when it is added before (Fig. 2C, no wash) or at the time of virus inoculation but is virtually inactive if it is added even shortly (<2 h) thereafter. In addition, the antiviral activity of the compound is severely attenuated if the compound is washed away before virus inoculation (Fig. 2C, wash), indicating that the compound must be present at the time of inoculation for optimal antiviral activity. This narrow time-of-addition window indicates that compound 2 targets a very early step in the viral life cycle.
Since the compounds fail to control infection if they are added shortly after inoculation, they should not have antiviral activity in persistently HCV-infected cells. To test this hypothesis, persistently infected cells were treated with compound 1 or 2 (5 μM) for 48 h. Parallel cultures were treated with a well-characterized polymerase inhibitor (2′-C-m-adenosine [2mAde]) as a positive control (9). As shown in Fig. 2D, compounds 1 and 2 had little or no effect on intracellular HCV RNA content or intracellular- or extracellular-infectivity titers. In contrast, 2mAde inhibited all of these parameters (Fig. 2D). These results demonstrate that compounds 1 and 2 have limited or no antiviral activity against established HCV infection. They also suggest that the compounds do not inhibit HCV RNA replication per se and that the ability of the compounds to suppress the accumulation of intracellular HCV RNA after inoculation at a high MOI (Fig. 2A) likely reflects their ability to inhibit a step preceding steady-state HCV RNA replication.
Similar single-cycle-infection experiments with HCV and other RNA viruses, such as lymphocytic choriomeningitis virus (LCMV) and Borna disease virus (BDV), were conducted in parallel using Huh-7 cells. Figure 3 shows that antiviral doses of compound 1 and 2 have little or no impact on LCMV or BDV infection efficiency, indicating that they display specific anti-HCV activity.
Since compounds 1 and 2 interfered with early events of HCV infection without affecting persistent HCV infection, we asked whether they blocked viral entry by using HCV-pseudotyped retroviral particles that recapitulate envelope glycoprotein-mediated receptor binding, particle internalization, and low-pH-mediated membrane fusion (3). Figure 4 shows that neither compound (5 μM) interfered significantly with either HCVpp or VSVpp infection, in contrast with the entry inhibitor fluphenazine (5 μM), which selectively reduced HCVpp infection by at least 1 order of magnitude without interfering with VSVpp infection, as previously described (22). These results suggest that the compounds interfere with a step downstream of glycoprotein-mediated fusion.
We examined the impact of compounds 1 and 2 on the ability of subgenomic HCV RNAs to initiate RNA replication after transfection into naïve target cells. This recapitulates the early steps of HCV infection after the viral genome is released into the cytoplasm. Huh-7 cells were transfected with a JFH-1 subgenomic replicon bearing a luciferase reporter gene in the absence or presence of compound 1 or 2 (5 μM) or the polymerase inhibitor 2mAde (5 μM), and luciferase activity was measured in total cell lysates at various times postinfection, as shown in Fig. 5A. Luciferase activities were comparable under all conditions at 5 and 12 h posttransfection, indicating that compounds 1 and 2 do not interfere with primary HCV internal ribosome entry site (IRES)-dependent translation of the incoming HCV RNA. This is underscored by the comparable luciferase activities observed in cells transfected with a replication-deficient mutant replicon and treated with compound 1 or 2 or 2mAde (Fig. 5B).
Luciferase activity increased in the transfected, DMSO-treated cells 24 and 48 h after transfection but not in the 2mAde-treated cells (Fig. 5A) or in the cells transfected with the replication-deficient replicon (Fig. 5B), indicating that luciferase activity reflected active HCV RNA replication. During the initial replication phase, from 12 to 24 h posttransfection, luciferase activity was slightly reduced (~3-fold) by compound 1 and more strongly reduced (~20-fold) in compound 2-treated cells (Fig. 5A). These results indicate that compounds 1 and 2 inhibit the initiation of HCV RNA replication, at a step downstream of primary translation, and that they do so proportionally to the antiviral activity they display in single-cycle infections. Interestingly, even though fresh compound was added to the cultures 24 h posttransfection, replicon replication between 24 and 48 h was inhibited only minimally in compound 1-treated cells and only moderately in compound 2-treated cells. This was in clear contrast with the profound antiviral activity of the polymerase inhibitor 2mAde (Fig. 5A). These results suggest that once HCV RNA replication is initiated, compound 1 and compound 2 are ineffective, implying that they address a unique step in the virus life cycle that is limited to the initiation phase of viral RNA replication.
To test this hypothesis, we treated the transfected cells with DMSO, the polymerase inhibitor 2mAde, or compound 1 or 2 at the time of transfection (0 h), 24 h later, or at both time points, and we measured luciferase activity 5, 24, and 48 h later (Fig. 5C, inset). As expected, administration of neither 2mAde nor compound 1 or 2 inhibited luciferase activity 5 h after transfection (prior to the onset of replication), but all agents inhibited luciferase activity 24 h after transfection (during which initiation of viral RNA replication occurs) (Fig. 5C). Inhibition was also observed at 48 h in cells that had been treated from the beginning of the experiment with compound 1 or 2, albeit slightly less than that observed at 24 h (Fig. 5C). However, when treatment was delayed until 24 h after transfection, neither compound 1 nor compound 2 displayed any antiviral activity, while the polymerase inhibitor was strongly suppressive (Fig. 5C), confirming that compounds 1 and 2 do not inhibit ongoing replication of HCV RNA. These results illustrate that compounds 1 and 2 inhibit the initiation of HCV RNA replication, thereby confirming the impact of these compounds in single-cycle-infection experiments and suggesting that they inhibit viral infection by interfering with the initiation of viral replication after translation of the incoming viral RNAs.
Similar transfection experiments were performed with a genotype 1b, Con1-based subgenomic replicon in Huh-7.5.1 cells. As expected based on this replicon's poor replicative capacity relative to that of JFH-1 (34), Con1-luc transfection induced luciferase activity that rapidly decreased after electroporation (Fig. 6A), while luciferase activity increased exponentially in JFH-1–luc–transfected cells (Fig. 6B). Nevertheless, Con1-luc replication was demonstrated by its sensitivity to the polymerase inhibitor 2mAde (Fig. 6A), similar to what has been reported using a replication-deficient mutant (34). However, because of the limited replication capacity of the Con1 replicon (Fig. 6A), the luciferase signal obtained at 24 h derived predominantly from translation of incoming genomes rather than from RNA replication, as shown by the similarity of the signals in the 2mAde-treated and untreated cells. Thus, using 2mAde-treated cells as a reference, we evaluated the antiviral activity of these compounds against establishment of Con1 RNA replication at 24 h, the earliest time point at which replication could be demonstrated (Fig. 6A). Figure 6C shows that compound 1 (5 μM) produced a small although significant (P < 0.05) reduction in luciferase accumulation at 24 h posttransfection and that compound 2 inhibited Con1 RNA replication with a magnitude comparable to that of the control inhibitor 2mAde (5 μM). These results suggest that this genotype 1b replicon is partially susceptible to the antiviral activity of these compounds, that compound 2 displays enhanced antiviral activity compared to that of compound 1, and that antiviral activity of this family of compounds is not limited to JFH-1.
In order to gain insight into the mode of action of compounds 1 and 2, we attempted to select JFH-1 variants that were resistant to the antiviral action of compound 2. We serially passaged D183 JFH-1 virus in the presence of sequentially increasing concentrations of compound 2. This led to the emergence of a compound 2-resistant (C2R) virus population; to reduce its infectivity by 50%, this virus required a concentration of compound 2 >10 fold higher than that for a control virus which was passaged in parallel in the absence of compound 2 (Fig. 7A). This reduced susceptibility was specific for compound 2, as the C2R virus displayed a sensitivity to 2mAde similar to that of the control virus (Fig. 7B). As shown in Fig. 7C, antiviral doses of compound 2 could not prevent intracellular HCV RNA accumulation in Huh-7 cells inoculated with the C2R variant at a high MOI, in contrast to the strong antiviral activity displayed against the control virus. These results confirmed that a resistant variant had been selected. In order to analyze the genetic determinants conferring resistance and therefore investigate which areas of the viral genome undergo selective pressure by compound 2, we sequenced the viral genomes present in the supernatants of three nonconsecutive passages of the virus under selection as well as control viruses passaged in the absence of compound. Common mutations were found in both control and C2R virus (Fig. 7D). Figure 7E summarizes the mutations found in all three passages of the C2R virus that were not found in the control supernatants and therefore might confer resistance to compound 2.
Nonsynonymous mutations were found in the coding sequences corresponding to E1 (V223D), E2 (N410D), p7 (H781Y), NS2 (Q931R), NS5A (F2004L and E2371A), and NS5B (K2493R and A2780T). Despite these genetic differences, control and C2R viruses display comparable growth curves (Fig. 8A). Since our results indicate that the antiviral activity could be recapitulated in the context of the subgenomic replicon (Fig. 5) that lacks the sequences encoding the structural proteins p7 and NS2, we focused our attention on the mutations in NS5A and NS5B. Individual, double, and quadruple mutations in NS5A and NS5B were engineered into a wild-type JFH-1 subgenomic replicon bearing a luciferase reporter gene. In vitro-transcribed RNAs (5 μg) were electroporated into Huh-7 cells in the presence of DMSO or antiviral doses (2.5 and 5 μM) of compound 2. Cell lysates were prepared at 5 and 24 h posttransfection and assayed for luciferase activity. All mutant replicons displayed comparable (<2-fold difference) replication capacities as measured by the luciferase signal obtained 24 h posttransfection in the absence of inhibitors (Fig. 8B). Figure 8C shows the expected dose-dependent reduction of luciferase signal for the wild-type replicon. The antiviral activity of compound 2 was clearly reduced against the replicons bearing all four mutations, the NS5A double mutation, and the single mutation F2004L, which were insensitive to 2.5 μM compound 2 and only partially inhibited at 5 μM (Fig. 7). In contrast, the NS5A E2371A mutant and both NS5B mutants displayed sensitivities similar to that of the wild type. These results indicate that the F2004L mutation (position F28 in NS5A), which is not conserved among different HCV genotypes (Fig. 9), is sufficient to confer partial resistance to the antiviral activity of compound 2 in the context of JFH-1 and suggest that NS5A might be directly or indirectly involved in the process targeted by these compounds.
Study of the HCV life cycle in surrogate models of infection, e.g., HCVpp and replicon cell lines, has permitted analysis of the cellular and molecular events that govern viral entry and steady-state HCV RNA replication. Aspects of the viral life cycle not recapitulated in these systems, such as those occurring downstream of endosomal internalization of the infectious particles, leading to primary translation of the viral genome and establishment of the replication complexes, are beginning to be characterized (5, 12) thanks to the development of cell culture infection systems that recapitulate every step of the HCV life cycle (37, 61, 64).
By adaptation of this cell culture system into an unbiased cell-based screening methodology (22), interrogation of chemical libraries for antiviral molecules against HCV enables discovery of antiviral compounds that interrupt aspects of the infection that were previously uncharacterized. Thus, this methodology has the potential not only to provide novel lead compounds for therapy against this important pathogen but also to identify novel chemical probes to study aspects of HCV infection that have been functionally and biochemically uncharacterized.
Using this methodology, we have identified and characterized a novel family of small molecules that efficiently inhibit viral spread at nanomolar concentrations. These compounds efficiently inhibited HCV infection without altering viral entry, as shown by the lack of antiviral activity against HCVpp (Fig. 4), or interfering with persistent HCV RNA replication and virus particle production in persistently infected cells (Fig. 2D).
Time-of-addition experiments indicate that the inhibition occurs at early steps of the infection, since addition of the compounds as early as 2 h after virus inoculation resulted in the virtual loss of antiviral activity (Fig. 2C). The results shown in Fig. 2C argue against a rapid loss of antiviral activity due to compound instability since the compounds retained full antiviral activity even when they were added to the cultures 6 h before virus inoculation. Since HCVpp entry was unaffected by these compounds, we postulated that they target steps downstream of viral entry. This hypothesis was confirmed by the antiviral activity of the compounds against the initiation of HCV RNA replication by HCV subgenomic replicons that were electroporated into naïve Huh-7 cells (Fig. 5A and B). This effect was dependent on the time of addition of the compounds and was not observed once HCV RNA replication had been initiated (Fig. 5C), recapitulating the lack of antiviral activity observed in persistently infected cells. Overall, these observations indicate that compounds 1 and 2 specifically target a very early step in HCV RNA replication without altering steady-state HCV RNA production. These results suggest that the compounds target a transient event that is rate limiting for the initiation of HCV RNA replication, rather than HCV RNA replication per se.
It has recently been shown that the cellular autophagy machinery is required for translation of incoming HCV genomes but not for translation of progeny genomes (12). Collectively, those observations and the results presented herein suggest that the cellular and viral factors required to initiate viral replication are different from those required to maintain it once replication complexes have formed. Another possible interpretation of our results derives from the fact that expression of HCV proteins, notably NS4B, causes a profound reorganization of cellular membranous compartments to promote the formation of replication complexes in modified ER membranes (5, 13). It is therefore formally possible that the compounds described above target HCV RNA replication per se but that virus-induced cellular changes (e.g., alteration of intracellular membranes) reduce the access of the compounds to their molecular target(s) sufficiently once replication complexes have been established to reduce their efficacy, resulting in an apparent lack of antiviral activity.
Selection of virus variants with reduced susceptibility to compound 2 (C2R virus) demonstrates the specific antiviral activity of this compound. We have shown that a point mutation in NS5A (F2004L) identified in the resistant virus variants is sufficient to confer the resistance phenotype to subgenomic replicons (Fig. 8), suggesting that the mechanisms by which these compounds inhibit HCV infection and establishment of the subgenomic-replicon replication are the same. Our results also indicate that mutation F2004L did not result in a significant fitness cost, as the C2R virus grows to titers comparable in magnitude (106 FFU/ml) and kinetics to those of the control virus (Fig. 8A). Moreover, introduction of the F2004L point mutation into the subgenomic replicon did not impair baseline replication (Fig. 8B), indicating that this mutation does not weaken viral replication even in a wild-type genetic background.
The NS5A residue F28 (F2004 in the polyprotein) is not conserved among all genotypes and strains, e.g., Con1 and J4L6 from genotype 1b display a leucine residue in that position (Fig. 9). However, a Con1-based replicon was susceptible to compound 2 (Fig. 6C) even at a concentration (2.5 μM) that was inactive against the JFH-1 F28L mutant, suggesting that a leucine residue in position 28 does not confer resistance to compound 2 in the context of genotype 1b. In any case, in the context of JFH-1 (genotype 2a), position F28 is located in a conserved class I proline-rich (PR) motif immediately adjacent to the membrane-anchoring amphipathic alpha helix located at the N terminus of NS5A (57). Although no specific function has been attributed to this conserved PR motif, PR motifs in NS5A have been proposed as structural elements that determine interaction of NS5A with cellular factors containing Src homology 3 (SH3) domains (57). This could be demonstrated for the C-terminal proline-rich motif, which mediates interaction with kinases like phosphatidylinositol 3-kinase (PI3K) (56), Lck, Hck, and Fyn (40), as well as adaptor proteins such as Grb2 (57) and amphiphysin II/BinI (63), but not for the N-terminal motif, in which the mutation conferring resistance to compound 2 (F28L) is included. It is therefore possible that compounds of the chemical family described here interfere with interactions of NS5A via its N-terminal PR motif and that mutation F28L restores such interactions that could be essential for the establishment of HCV RNA replication. Position F28 is located in close proximity to the N-terminal membrane-anchoring domain in NS5A and might contribute to the interaction of this protein with cellular membranes. It is possible that, during the initial phases of the infection, compounds 1 and 2 interfere with docking of NS5A into membranes devoid of viral replication complexes and that this does not occur once replication complexes have been formed. While these speculative scenarios would explain our observations, extensive biochemical studies are required to define the molecular events underlying both antiviral activity and resistance. Studies aimed at determining the composition of the molecular complexes that compounds 1 and 2 may bind in the cell will contribute to understanding the mechanisms underlying the establishment of HCV RNA replication complexes, a process in which NS5A could play a specific role.
NS5A has recently been revealed as a novel target for potent antiviral compounds that inhibit viral RNA replication (e.g., BMS-790052) (17). Genetic evidence indicates that compound 2 targets NS5A directly or indirectly (Fig. 6) but seems to do so by a different, unprecedented mechanism that targets initiation but not steady-state HCV RNA replication. Remarkably, despite this apparent differential mode of action, position F28 in JFH-1 is aligned with position 28 in genotypes 1a and 1b (Fig. 9), where mutations M28T and L28T, respectively, have been reported to confer partial resistance to BMS-790052 (14, 17). These genetic data indicate that both compounds impose similar selective pressures on the N terminus of NS5A, suggesting that they could target similar functions in NS5A but appear to do so by different molecular mechanisms.
Multiple functions involving different steps in the viral life cycle have been attributed to NS5A. However, no specific enzymatic activity has been ascribed to NS5A, and the molecular mechanisms by which it functions remain elusive. Thus, the inhibitors described herein constitute potentially powerful chemical tools that could facilitate our understanding of the functions that NS5A plays in the viral life cycle and may lead to the development of novel compounds with improved therapeutic potential.
We are grateful to Takaji Wakita (National Institute of Infectious Diseases, Tokyo, Japan) for providing the infectious JFH-1 molecular clone and replicon constructs, Charles Rice (Rockefeller University) for providing the Huh-7.5.1 cells from which the Huh-7.5.1-c2 cells were derived, Mansun Law and Dennis Burton (The Scripps Research Institute, La Jolla, CA) for providing the recombinant human IgG anti-E2, François-Loïc Cosset (Ecole Normale Superieure, INSERM U758, Lyon, France) for providing the vectors necessary for HCVpp production, Jeffrey Glenn from Stanford University (Palo Alto, CA) for providing the Con1-luc replicon, and Weidong Zhong from Gilead Sciences (Foster City, CA) for providing the 2′-C-m-adenosine. We thank our colleagues Urtzi Garaigorta and Stefan Wieland for their expert advice and useful discussions. We are grateful to Brian Boyd for excellent technical assistance.
This work was supported by NIH grants R01-CA108304 and R01-AI079043.
This is manuscript number 20959 from The Scripps Research Institute.
Published ahead of print on 23 March 2011.