Inhibition of Hepatitis C Virus RNA Replication by 2′-Modified Nucleoside Analogs*

The RNA-dependent RNA polymerase (NS5B) of hepatitis C virus (HCV) is essential for the replication of viral RNA and thus constitutes a valid target for the chemotherapeutic intervention of HCV infection. In this report, we describe the identification of 2′-substituted nucleosides as inhibitors of HCV replication. The 5′-triphosphates of 2′-C-methyladenosine and 2′-O-methylcytidine are found to inhibit NS5B-catalyzed RNA synthesis in vitro, in a manner that is competitive with substrate nucleoside triphosphate. NS5B is able to incorporate either nucleotide analog into RNA as determined with gel-based incorporation assays but is impaired in its ability to extend the incorporated analog by addition of the next nucleotide. In a subgenomic replicon cell line, 2-C-methyladenosine and 2′-O-methylcytidine inhibit HCV RNA replication. The 5′-triphosphates of both nucleosides are detected intracellularly following addition of the nucleosides to the media. However, significantly higher concentrations of 2′-C-methyladenosine triphosphate than 2′-O-methylcytidine triphosphate are detected, consistent with the greater potency of 2′-C-methyladenosine in the replicon assay, despite similar inhibition of NS5B by the triphosphates in the in vitroenzyme assays. Thus, the 2′-modifications of natural substrate nucleosides transform these molecules into potent inhibitors of HCV replication.


Hepatitis C virus (HCV)
infection is the leading cause of sporadic, post-transfusion, non-A non-B hepatitis (1,2). One hundred seventy million people worldwide are thought to be infected with hepatitis C virus of which an estimated 4 million reside in the United States (3). Approximately 80% of infected individuals progress to chronic infection. Long term chronic HCV infection can lead to liver cirrhosis and to hepatocellular carcinoma (4 -6). Currently, the recommended therapy is treatment with a combination of interferon ␣2b and ribavirin, which results in a sustained viral response in 40% of patients (7,8). Investigational therapies using a combination of pegylated interferon and ribavirin have lead to an sustained viral response in 54% of patients, but the response rate (42%) of patients harboring HCV genotype 1 is lower (9,10). Consequently, additional therapies for HCV infection are needed.
Antiviral chemotherapies based on administration of analogs of deoxynucleosides have been widely successful as treatment for HIV, herpes virus, and hepatitis B infection (11,12). Intracellular phosphorylation of the nucleoside analog to the triphosphate creates the active form of the inhibitor that then serves as a substrate for the viral polymerase. Generally, incorporation of the nucleotide analog at the 3Ј-end of the replicating viral DNA causes termination of DNA synthesis, owing to the lack of the 3Ј-hydroxyl required for extension. These successes suggest that an investigation of ribonucleoside analogs as inhibitors of HCV replication would be worthwhile.
The HCV NS5B protein, the RNA-dependent polymerase responsible for the synthesis of the viral RNA genome, is an attractive target for the development of antiviral agents (13). The enzymatic activity of NS5B has been extensively characterized in vitro (13)(14)(15)29). Additionally, cell lines that harbor subgenomic replicons capable of supporting HCV replication are available to assess inhibition of replication by compounds within the cellular environment (16,17). The antiviral effect of interferon ␣ has been documented in these lines (18).
Screens of available nucleosides for inhibitors in the cellbased bicistronic replicon assay have identified two nucleoside analogs, 2Ј-C-methyladenosine and 2Ј-O-methylcytidine, that specifically inhibit HCV RNA replication in the absence of cytotoxicity. The biochemical basis for the inhibition by these nucleoside analogs has been investigated. When added to replicon cells growing in culture, the nucleoside analogs resulted in the intracellular formation of the corresponding triphosphates that were shown to be potent, competitive inhibitors of NS5B-catalyzed reactions in vitro. This study demonstrates the utility of 2Ј-substituted nucleosides in the inhibition of HCV RNA replication. * The use of the compounds 2Ј-C-methyladenosine and 2Ј-O-methylcytidine as therapies for HCV infection is disclosed in international patent applications WO 01/90121 A2 assigned to Idenix, Inc. and WO 02/57425 A2 assigned to Merck & Co., Inc. and Isis Pharmaceuticals, Inc. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
2Ј-C-Methyladenosine triphosphate was synthesized according to the general procedures previously described (21). The triphosphate was purified by anion exchange chromatography using a 30-ϫ 100-mm Mono Q column (Amersham Biosciences) with a buffer system of 50 mM Tris, pH 8. The elution gradient was 40 mM to 0.8 M NaCl in two column volumes. Appropriate fractions from Mono Q chromatography were collected and desalted by reverse-phase (RP) chromatography using a Luna C18 250-ϫ 21-mm column (Phenomenex) with an elution gradient from 1% to 95% methanol in 5 mM triethylammonium acetate. Mass spectra of the purified triphosphate were determined using in-line RP HPLC mass spectrometry on a Hewlett-Packard (Palo Alto, CA) MSD 1100. The molecular mass was determined using the Hewlett-Packard Chemstation analysis package. LC/MS: 520.0 (calc. for C 11 H 17 N 5 O 13 P 3 : 520.0036). The purity of the nucleoside triphosphate was determined with analytical RP and anion exchange HPLC to be 100%.
NS5B Enzyme Assay on Template t500 -RNA polymerase activity was determined in reactions catalyzed by NS5B⌬21 and NS5B⌬55 by measuring the incorporation of radiolabeled NTPs into a heteromeric RNA template via a copy-back mechanism. Template t500 was generated by T7 runoff transcription as previously described (22), using commercially available kits (Ambion) following the manufacturer's instructions. The t500 was purified with RNeasy kits (Qiagen) and was quantified using absorbance at 260 nm. NS5B-catalyzed reaction conditions included 500 nM NS5B⌬21 or 25 nM NS5B⌬55 in a 50-l reaction containing 0.75 g of t500, 20 mM Tris, pH 7.5, 80 mM KCl, 2 mM MgCl 2 , 5 mM DTT, 0.4 unit/l RNasin (Promega), 0.2% polyethylene glycol 8000, 50 M EDTA, 1 M NTPs (unless otherwise noted), and ϳ1 Ci of either [␣-32 P]-or [␣-33 P]GTP or -ATP. Reactions were initiated by the addition of a mixture of NTPs, after preincubation of the other reaction components for 30 min at RT. Reactions proceeded for 2 h at RT and were quenched by addition of EDTA. Product formation was determined by DE-81 filter binding (Whatman) as previously described (22,24). The inhibitor concentration at which the enzyme-catalyzed rate is reduced by half (IC 50 ) was determined by fitting the relative rate data to the Hill equation, where v i is the reaction velocity in the presence of inhibitor, v 0 is the reaction velocity in the absence of inhibitor, and n is the Hill coefficient.
Data fitting was carried out with use of Kaleidagraph (Synergy Software, Reading, PA). Gel-based Incorporation Assay-The incorporation into synthetic RNA and extension of nucleoside analogs catalyzed by HCV NS5B was determined in reactions utilizing 5Ј-end-labeled oligoribonucleotides. The incorporation/extension of analogs of adenosine triphosphate was analyzed in reactions on template 68N (sequence 5Ј-AGAUGGCCCG-GUUUUCCGGGCC-3Ј), which is designed to fold into a hairpin structure with the first available template base as a U (underlined in the sequence). The concentration of oligonucleotide 68N was determined by absorbance at 260 nm and the oligonucleotide was 5Ј-end-labeled with [␥-32 P]ATP in reactions catalyzed by T4 polynucleotide kinase (US Biochemicals, Cleveland, OH or Invitrogen, Gaithersburg, MD) as previously described (25). NS5B⌬55 (1 M) was preincubated with template 68N (600 nM) in reaction buffer containing 20 mM Tris, pH 7.5, 80 mM KCl, 5 mM DTT, 50 M EDTA, 2 mM MgCl 2 , 0.4 unit/l RNasin (Promega) for 30 min at room temperature. Reactions were initiated by the addition of NTPs. Reactions included 10 M ATP and/or 2 M UTP or 1-50 M 2Ј-C-methyladenosine triphosphate with or without 2 M UTP. Reactions were allowed to proceed for 30 or 60 min, and then 5-l aliquots were removed and quenched with 15 l of formamide gel load buffer. After denaturing the RNA at 65°C for 30 min, substrate and one or more products were separated on 20% acrylamide-8 M urea gels and analyzed using a PhosphorImager (Amersham Biosciences). In a similar manner analogs of cytidine triphosphate were examined for incorporation/extension in reactions with oligonucleotide 76N (sequence 5Ј-ACUGGGCCCGGUUUUCCGGGCC-3Ј). Oligoribonucleotides 68N and 76N were synthesized using 2Ј-acetate ester chemistry (Dharmacon, Lafayette, CO), purified using denaturing PAGE gels, and deprotected according to the manufacturer's instructions.
In Situ Ribonuclease Protection Assay-HBI10A cells (27) were grown and assayed as previously described. 2 Replicon cells were passaged at 1:5 and plated at a cell density of 40,000 cells/well in Cytostar plates (Amersham Biosciences) in complete Dulbecco's modified Eagle's medium media containing 10% FBS and 0.8 mg/ml G418. Compound dissolved in Me 2 SO was added to the cells at a final Me 2 SO concentration of 1% and incubated for 24 h at 37°C/5% CO 2 . Cells were fixed in 10% formalin/phosphate-buffered saline, permeabilized in 0.25% Triton X-100, and then hybridized with an antisense 33 P-labeled RNA probe (1.0 -2.0 ϫ 10 4 cpm/l) in formamide hybridization buffer (Amersham Biosciences) at 50°C overnight. The RNA probe was generated with T7 runoff transcription and had a sequence complementary to nucleotides 1184 -1481 of the NS5B gene. Plates were treated with 20 g/ml RNase A at room temperature for 30 min, washed with 0.25ϫ SSC buffer at room temperature and then at 65°C for 20 min each wash, and then counted in a TopCount plate reader.
Cytotoxicity Assay-Cytotoxicity was assayed as previously described. 2 The cells were plated in 96-well plates in parallel at the same density, and the compound was added as above. At the indicated times, Promega Cell Titer 96 Aqueous One Solution Reagent (MTS) was added for 1 h at 37°C/5% CO 2 and then absorbance was read at 490 nm in a plate reader.
Intracellular Metabolism Studies-Two cell lines, Huh-7 and HBI10A, were used for intracellular metabolism studies of 5-[ 3 H]-2Ј-Omethylcytidine and 7-[ 3 H]-2Ј-C-methyladenosine. Huh-7 is a human hepatoma cell line, and HBI10A denotes a clonal line derived from Huh-7 cells that harbors the HCV bicistronic replicon. Huh-7 cells were plated in complete Dulbecco's modified Eagle's medium containing 10% fetal bovine serum and HBI10A cells in the same containing G418 (0.8 mg/ml) at 1.5 ϫ 10 6 cells/60-mm dish such that cells were 80% confluent at the time of compound addition. Tritiated compound was incubated at 2 M in the cell medium for 3 or 23 h. Cells were collected, washed with phosphate-buffered saline, and counted. The cells were then extracted in 70% methanol, 20 mM EDTA, 20 mM EGTA, and centrifuged. The lysate was dried, and radiolabeled nucleotides were analyzed using an ion-pair reverse phase (C-18) HPLC on a Waters Millenium system connected to an in-line ␤-RAM scintillation detector (IN/US Systems). The HPLC mobile phases consisted of (a) 10 mM potassium phosphate with 2 mM tetrabutylammonium hydroxide and (b) 50% methanol containing 10 mM potassium phosphate with 2 mM tetrabutylammonium hydroxide. Peak identification was made by comparison of retention times to standards.

RESULTS
Inhibition of NS5B Enzyme Activity-NS5B-catalyzed incorporation of nucleotides in reactions with template t500 generates a copy-back or hairpin product (22), as previously described for other RNA templates (28). The rate of product formation catalyzed by NS5B⌬21 or NS5B⌬55 was reduced in the presence of either 2Ј-C-methyladenosine triphosphate or 2Ј-O-methylcytidine triphosphate (structures shown in Fig. 1), with IC 50 values as shown in Table I, as determined by monitoring the total incorporation of radiolabeled nucleotide as described under "Experimental Procedures." The Hill coefficients did not significantly differ from unity. The potency of inhibition by either nucleotide analog was not affected whether the radiolabeled nucleoside triphosphate was GTP or ATP (data not shown), indicating that replacement of the radiolabel by the nucleoside analog was not responsible for the inhibition.
Mode of Inhibition-To determine the mode of inhibition by 2Ј-C-methyladenosine triphosphate and 2Ј-O-methylcytidine triphosphate of RNA synthesis catalyzed by NS5B⌬21, reactions were run in which the concentrations of ATP and CTP were varied, respectively, holding the concentrations of the other NTPs constant at concentrations above K m . Double-reciprocal plots of the data as shown in Fig. 2 indicated competitive inhibition of activity by 2Ј-C-methyladenosine triphosphate and 2Ј-O-methylcytidine triphosphate with varying ATP and CTP, respectively. The K i value as determined from a replot of the slopes of the double-reciprocal plot was 0.9 M for 2Ј-C-methyladenosine triphosphate and 0.3 M for 2Ј-O-methylcytidine triphosphate.
Gel-based Incorporation Assay-To determine whether NS5B⌬55 is capable of incorporating the nucleoside analogs into a growing RNA strand, gel-based analyses of reactions using hairpin RNA templates were carried out. The sequence of the RNA template (68N) was designed such that an intramolecular hairpin could form, allowing the incorporation of AMP followed by UMP. NS5B⌬55 showed a much greater ability to incorporate nucleotides efficiently onto the hairpin RNA substrates than did NS5B⌬21. The relatively low activity of NS5B⌬21 with the hairpin templates is likely a consequence of the low fraction of catalytically competent NS5B⌬21 (ϳ2%, Ref. 22) compared with NS5B⌬55 (ϳ40%, data not shown). The incorporation of AMP leads to the appearance of a single product band (Fig. 3A, lane 3). In reactions that included ATP and UTP, the product resulting from the incorporation of AMP was completely extended by addition of UMP (Fig. 3A, lane 4). NS5B⌬55 was capable of incorporating 2Ј-C-methyladenosine monophosphate onto the 3Ј-end of the RNA at the lowest nucleotide concentration tested, 1 M. However, as shown in Fig.  3A (lanes 9 -11), NS5B⌬55 could not add uridine monophosphate onto the 2Ј-C-methyladenosine-terminated template efficiently, although a trace amount of extended product was evident. In a similar manner, NS5B⌬55 was capable of incorporating 2Ј-O-methylcytidine monophosphate onto the 3Ј-end of RNA hairpin 76N as shown in Fig. 3B. However, no detectable extension product was visible when the next correct nucleoside triphosphate, ATP, was added to the reaction.
Inhibition of DNA Polymerases ␣, ␤, and ␥-To determine the specificity of inhibition the activity of human DNA polymerases ␣, ␤, and ␥ were monitored in vitro in the presence of  Nucleosides 2Ј-C-Methyladenosine and 2Ј-O-Methylcytidine Inhibit HCV RNA Replication in Cells-Nucleosides 2Ј-Cmethyladenosine and 2Ј-O-methylcytidine were tested for inhibitory activity in a cell-based replicon assay using a stable Huh-7 human hepatoma cell line, which supports the replication of HCV RNA and proteins. The effect of the nucleosides upon RNA replication in a clonal line designated HBI10A (27) was detected by in situ ribonuclease protection assay as previously described. 2 Representative titrations of the compounds in the replicon assay are shown in Fig. 4. Both compounds were active in the assay at 24 h with IC 50 values of 0.3 M for the 2Ј-C-methyladenosine and 21 M for the 2Ј-O-methylcytidine (Table II). The antiviral activity of both compounds was observed in the absence of cytotoxicity in HBI10A cells as measured in the MTS assay when tested up to 100 M.
Intracellular Metabolism to the Active Triphosphate-The intracellular metabolism of the tritiated versions of 2Ј-Cmethyladenosine and 2Ј-O-methylcytidine was studied in Huh-7 and HBI10A cells. The compounds were incubated for 3 or 23 h at 2 M prior to extraction and HPLC analysis, as shown in Fig. 5. The results are summarized in Table III. 2Ј-C-Methyladenosine was efficiently taken up into the cells and converted to its corresponding triphosphate. In contrast, incubation of HBI10A cells with 2Ј-O-methylcytidine yielded very little triphosphate and suffered from extensive metabolism to species that are consistent with (by comparison to retention times for appropriate controls) UTP and CTP, indicating that this molecule was subject to deamination and base swapping in the cell.

DISCUSSION
The advent of the cell-based, subgenomic, bicistronic replicon system (16) as a means of assessing the replication of viral a Compounds were incubated in cell culture for 24 h prior to determination of the relative amount of HCV replicon RNA with the in situ ribonuclease protection assay (27,28). Data were fit to Equation 1 to determine IC 50 values, and the values are the mean Ϯ S.D. from at least three independent experiments.
b Compound cytotoxicity was determined by MTS assay on parallel samples at the same time.
RNA within the cellular environment has permitted the evaluation of analogs of ribonucleosides to complement ongoing efforts aimed at identifying inhibitors of purified NS5B in vitro. Screening of available nucleosides for inhibition of viral replication in the replicon assay identified two inhibitory compounds, 2Ј-C-methyladenosine and 2Ј-O-methylcytidine.
The triphosphates of 2Ј-C-methyladenosine and 2Ј-O-methylcytidine inhibit the catalytic activity of purified HCV RNA polymerase with IC 50 values of 1.9 and 3.8 M, respectively ( Table I). Forms of HCV RNA polymerase having two different C-terminal truncations that were investigated have significantly different catalytic efficiencies, with NS5B⌬55 having ϳ20-fold greater specific activity than NS5B⌬21. However, IC 50 values varied only slightly between the two enzyme forms for the nucleoside analog triphosphates investigated. As expected the two nucleoside analog triphosphates were competitive inhibitors with varying nucleoside triphosphate, having K i values that were submicromolar.
Analysis of the incorporation of the nucleoside analogs onto a growing RNA strand was carried out using synthetic RNA templates that are designed to fold into intramolecular hairpins. NS5B⌬55 is capable of incorporating both 2Ј-C-methyladenosine monophosphate and 2Ј-O-methylcytidine monophosphate onto the appropriate RNA template, implying that both triphosphates can bind to the enzyme in the substrate NTP binding site and further implying there is some additional room in the vicinity of the 2Ј-carbon and the 2Ј-oxygen when bound in the active site that allows HCV NS5B to accommodate either the 2Ј-C-methyl or 2Ј-O-methyl substituent. The presence of 2Ј-substituents likely confers specificity of inhibition of the viral RNA polymerase over inhibition of the human DNA polymerases tested. After incorporation of the nucleotide analog, NS5B⌬55 is not capable of efficiently extending the incorporated analog by addition of the next correct nucleotide, suggesting that with this template system, the nucleotide analogs act as functional chain terminators, despite the presence in both cases of a 3Ј-OH. The results suggest that after incorporation the 3Ј-OH is not able to perform nucleophilic attack on the ␣-phosphorous of the incoming NTP. Further investigation is necessary to understand the molecular details of the apparent chain termination. Chain termination by incorporation of nucleotide analogs that retain a 3Ј-hydroxyl has previously been observed in the inhibition of DNA polymerase ␤ by arabinofuranosyladenine triphosphate (29) and arabinofuranosylcytidine triphosphate (30).
The nucleosides were converted intracellularly to the corresponding triphosphates, which, in turn, functioned as specific inhibitors of HCV RNA synthesis. The potency of inhibition of HCV replication observed in replicon-containing cells correlated with the levels of triphosphates formed intracellularly. A lesser amount of the intracellular 2Ј-O-methylcytidine triphosphate was detected than the 2Ј-C-methyladenosine triphosphate, and the difference is reflected in the reduced potency of 2Ј-O-methylcytidine in replicon-containing cells, despite the equivalent inhibition of the purified enzyme by the two triphosphates.
The potency of inhibition by 2Ј-C-methyladenosine in the cell-based replicon assay (0.3 M) is greater than the potency of the corresponding triphosphate in the enzyme assay (1.9 M). The greater potency in the cell-based assay likely reflects a combination of the high intracellular concentration of the corresponding triphosphate that is achieved (105 pmol/million cells in the presence of 2 M extracellular nucleoside) and the fact that the analog acts as a functional chain terminator. Once the nucleotide analog is incorporated into the replicon RNA, the resulting truncated RNA chain, in the absence of a known proofreading activity, is nonfunctional as a template for subsequent rounds of viral RNA synthesis. However, in the enzyme assay, once the nucleotide analog has been incorporated, the truncated RNA is still counted as the product.
Because the in vitro enzyme assay is performed under conditions that differ from physiological, it is conceivable that the enzyme assay may not be truly representative of biological activity. Nonetheless, the use of in vitro enzyme assays is validated by the demonstration of the inhibition of purified HCV NS5B by the triphosphates of nucleoside analogs that also are capable of inhibiting HCV replication in cell culture in the absence of cytotoxicity. A more definitive correlation of enzyme inhibition with antiviral effect in the replicon assay will be ascertained when resistant mutations are identified that confer resistance both in vitro and in the cell culture assay.
Whether the replicon assay will be a meaningful predictor of antiviral activity in vivo has yet to be determined. Equally uncertain at this point is the question of whether 2Ј-C-methyladenosine or 2Ј-O-methylcytidine have pharmacokinetic and safety profiles that are sufficiently attractive to warrant their development as HCV therapeutics. However, the current work establishes the direct inhibition of HCV RNA polymerase activity by 2Ј-modified nucleotides leading to inhibition of HCV replication in cells.