Identification of AP80978, a Novel Small-Molecule Inhibitor of Hepatitis C Virus Replication That Targets NS4B

Abstract

A small-molecule inhibitor of hepatitis C virus (HCV) designated AP89652 was identified by screening a compound library with an HCV genotype 1b subgenomic replicon assay. AP89652 contains two chiral centers, and testing of two syn enantiomers revealed that activity in the replicon assay resided with only one, AP80978, whose 50% effective concentration (EC₅₀) (the concentration at which a 50% reduction in Renilla luciferase levels was observed relative to an untreated control) was 630 nM. AP80978 was inhibitory against HCV genotypes 1a and 1b but not genotype 2a. In a replicon clearance assay, the potency and clearance rate of AP80978 were similar to those of telaprevir (VX950) and cyclosporine (CsA). AP80978 was nontoxic when tested against a panel of human cell lines, and inhibitory activity was HCV specific in that there was limited activity against negative-strand viruses, an alphavirus, and flaviviruses. By selection of resistant replicons and assessment of activity in genotype 1b/2a intergenotypic replicons, the viral protein target of this compound was identified as NS4B. NS4B F98V/L substitutions were confirmed by site-directed mutagenesis as AP80978 resistance-associated mutations. When tested against HCV produced in cell culture, the compound was significantly more potent than other HCV inhibitors, including VX950, CsA, and 2′-C-methyladenosine (2′C-meA). In addition, AP80977, the enantiomer that was inactive in the replicon assay, had activity against the virus, although it was lower than the activity of AP80978. These results suggest that AP80978 has the potential to be optimized into an effective antiviral drug and is a useful tool to further study the role of NS4B in HCV replication.

INTRODUCTION

Hepatitis C virus (HCV) is a positive-strand RNA virus belonging to the Flaviviridae family. Within the viral genome, the internal ribosome entry site (IRES) drives translation of a single polypeptide that is cleaved by both cellular peptidases and viral proteases to produce viral structural and nonstructural proteins (1). The virus-encoded RNA-dependent RNA polymerase NS5B is exceptionally error-prone, resulting in significant genome sequence variability. Based on sequence differences, HCV can be categorized into seven distinct genotypes, which differ both in global distribution and response to therapy (2, 3).

Worldwide, over 170 million people are infected with this virus (4). HCV infection is a major cause of chronic liver disease, such as cirrhosis and hepatocellular carcinoma, and is a leading cause of liver transplantation (5, 6). Until recently, the standard of care (SOC) was a combination of pegylated interferon and ribavirin, which is commonly associated with severe side effects and low sustained virological response rates for patients infected with HCV genotype 1, the most prevalent genotype in North America and Europe (7, 8). Direct-acting antivirals (DAA) have been the focus of intensive drug discovery efforts, particularly the viral NS3-4A protease, the NS5A phosphoprotein, and the NS5B polymerase. A triple combination composed of the SOC with one of two protease inhibitors, telaprevir (VX950) or boceprevir, enhances cure rates and is now approved for treatment of patients with chronic HCV genotype 1 infection (9, 10). However, resistance develops quickly to these and other antiviral compounds, and severe side effects and drug interactions complicate treatment (11). New protease and polymerase inhibitors have recently been approved, but the development of additional classes of antiviral compounds against novel viral targets will broaden treatment options and provide multiple options for interferon-free HCV therapy (1, 3, 12–14).

To this end, we carried out a high-throughput, cell-based HCV genotype 1b subgenomic replicon screen to identify novel compounds with antiviral activity against HCV. One compound that was selected for further study was a molecule with two chiral centers, designated AP89652. After separation of enantiomers, antiviral activity was found to be associated with AP80978, one of the two tested isomers. The active enantiomer was genotype 1 specific, noncytotoxic, and inactive against numerous other virus replication systems, which included other flaviviruses. Two approaches were taken to study the molecular target of the compound, including selection of resistant replicons and generating intergenotypic 1b/2a replicons and virus with differential susceptibility to the compound. Both approaches indicated a novel target of this compound, HCV NS4B. When tested against HCV produced in cell culture, the compound was significantly more potent than other HCV inhibitors, including VX950, cyclosporine (CsA), and 2′-C-methyladenosine (2′C-meA).

MATERIALS AND METHODS

Maintenance of Huh-7.5 cells.

Huh-7.5 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (FBS) (Invitrogen), 100 units/ml penicillin (Invitrogen), and 100 μg/ml streptomycin (Invitrogen) at 37°C in a humidified 5% CO₂ incubator. The cells were subcultured by washing them once with phosphate-buffered saline (PBS) (Invitrogen), followed by incubating them for up to 5 min in 0.05% trypsin–EDTA (Invitrogen) at 37°C until the cells detached from the vessel. Upon detachment, complete medium was added to inactivate trypsin, and cells were counted and seeded at the desired density into T-flasks (TPP; Midwest Scientific, St. Louis, MO). The cells were grown to 80% maximum confluence and seeded at a density of no less than 13,000 cells/cm².

Preparation of CA32 replicon cells.

The replicon utilized in the primary and secondary screens, CA32, was created at Apath, LLC, and is a transient HCV genotype 1b subgenomic replicon generated from the Con1 strain. In this replicon, the HCV internal ribosome entry site (IRES) within the 5′ nontranslated region (5′ NTR) drives translation of the first 32 amino acids of the core protein fused to humanized Renilla luciferase (hRluc). The encephalomyocarditis virus (EMCV) IRES lies 3′ of the hRluc open reading frame (ORF) and drives translation of nonstructural (NS) proteins NS3 through NS5B, which is flanked at its 3′ end with the HCV 3′ NTR (for a schematic, see Fig. 1A).

FIG 1.

(A) Schematic of HCV genotype 1b replicon (CA32) used in compound screening. (B) Structure of AP89652, with two chiral centers indicated with asterisks. (C) Structures of two syn enantiomers present within the AP89652 racemic mixture.

To prepare CA32 replicon cells, the replicon-encoding plasmid APP660 was linearized with the unique restriction endonuclease ScaI (NEB, Ipswich, MA) at 37°C for 4 h. The reaction was extracted once with phenol-chloroform-isoamyl alcohol (25:24:1) (Invitrogen) and once with chloroform-isoamyl alcohol (24:1). The linearized plasmid was precipitated with 1/10 volume of 3 M sodium acetate (pH 5.2) (Sigma-Aldrich, St. Louis, MO) and 2 volumes ethanol, and the pellet was washed twice with 70% ethanol. Linearized DNA (1 μg) was in vitro transcribed (IVT) using T7 polymerase from the MEGAscript T7 kit (Invitrogen). RNA size and integrity were verified by resolving 1 μg IVT RNA on a standard formaldehyde-agarose gel and staining with ethidium bromide (Sigma-Aldrich). Approximately 24 h prior to electroporation, Huh-7.5 cells were seeded at a density of 5.5 × 10⁶ cells per 150-cm² T-flask. On the day of electroporation, the cells were washed once with PBS prior to incubation with trypsin-EDTA for up to 5 min at 37°C. After the cells detached, complete medium was added, and the cells were pelleted via centrifugation at 200 × g for 5 min at 4°C. The cells were washed by resuspending them in ice-cold PBS and pelleting at 200 × g for 5 min at 4°C, resuspended once more in ice-cold PBS, passed through a 70-μm cell strainer (VWR, Radnor, PA), and counted prior to a final centrifugation at 200 × g for 5 min at 4°C. After resuspension in ice-cold PBS at a density of 1.5 × 10⁷ cells per ml, 6 × 10⁶ cells were combined with 1 μg IVT RNA and electroporated in a 2-mm-gap electroporation cuvette (Bio-Rad, Hercules, CA) using the BTX Electrosquare porator with the following settings: 820 V, 99-μs pulse length, 5 pulses, 1.1-s intervals. Following electroporation, the cells were recovered at room temperature for 10 min prior to plating in complete medium. Replicon-containing cells were passaged as they reached 70 to 80% confluence and were seeded at a density no lower than 13,000 cells/cm².

Compound screening with CA32 replicon cells.

Replicon cells were seeded into 96-well plates at 20,000 cells/well in screening medium, consisting of DMEM with 10% FBS, 1× penicillin-streptomycin, 1× nonessential amino acids, and 100 ng/ml Fungizone (Invitrogen). After 4-h attachment, the cells were treated with compound in 1% dimethyl sulfoxide (DMSO) (Sigma-Aldrich) in a final volume of 200 μl for 24 to 48 h. Pilot studies indicated that 1% DMSO caused no adverse effects on cell viability for this assay duration. For primary screening, compounds were added at a single dose (10 μM) in a single replicate without toxicity evaluation. Compounds that caused a 50% reduction in Renilla luciferase levels were progressed to secondary screening for 50% effective concentration (EC₅₀) (the concentration at which a 50% reduction in Renilla luciferase levels was observed relative to an untreated control) and cytotoxic concentration of drug that reduced the viable cell number by 50% (CC₅₀) determination. Identical culture conditions were used for EC₅₀ and CC₅₀ determination. For secondary screening, compound was added to wells using a five-point threefold serial dilution series with four replicate treatments per dose. Renilla luciferase levels were assayed using the Renilla luciferase assay kit (Promega, Madison, WI), and toxicity/cell viability was assayed using CellTiter-Glo luminescent cell viability assay (Promega). Both EC₅₀ and CC₅₀ values were determined from the raw data using a proprietary software program based on a Hill plot calculated from a four-parameter logistic model.

Compound synthesis and enantiomer separation.

AP89652 [3-chloro-5-(furan-2-yl)-N-(thiophene-2-yl-methyl)—7-(trifluoromethyl)-4,5,6,7-tetrahydropyrazolo[1,5,a]pyrimidine-2-carboxamide] (see Fig. 1B) was synthesized in two steps by using a procedure described by Dalinger et al. (15). The pyrazolo[1,5-a]pyrimidine ring was assembled by condensation of 5-amino-4-chloro-1H-pyrazole-3-carboxylic acid with 4,4,4-trifluoro-1-(furan-2-yl)-butane-1,3-dione, and the pyrimidine ring was reduced stereoselectively with sodium borohydride in ethanol, yielding only one pair of diastereoisomers that corresponds to 2,4-syn isomers. Separation of two enantiomers (AP80977 and AP80978) was achieved by a high-performance liquid chromatography (HPLC) method using a chiral column Chiralcel OD-H and heptane–2-propanol as a mobile phase. The absolute configuration of AP80978 was determined by X-ray diffraction (oXray Ltd.), and the active enantiomer was found to be (5S,7R)-3-chloro-5-(furan-2-yl)-N-(thiophene-2-yl-methyl)-7-(trifluoromethyl)-4,5,6,7-tetrahydropyrazolo [1,5,a]pyrimidine-2-carboxamide. The synthetic procedure for AP89652, its separation into two enantiomers, and determination of the absolute configuration for the active AP80978 enantiomer were described previously (16, 17).

Control inhibitors.

Control inhibitors ribavirin, a viral RNA synthesis inhibitor (18), and cyclosporine (CsA), an HCV replication inhibitor (19), were obtained from Sigma-Aldrich (St. Louis, MO). VX950, an NS3-4A protease inhibitor (10), was synthesized by Stereochem Research Centre (Hyderabad, India). 2′-C-methyladenosine (2′C-meA), an NS5B polymerase inhibitor (20), was obtained from Carbosynth Limited (Berkshire, United Kingdom).

Replicon clearance assay.

The genotype 1b replicon cells used in this assay harbor the replicon Con1/SG-Neo(I)hRLuc2aUb (contains the subgenomic replicon [SG]), which was created at Apath, LLC, and is encoded by plasmid APP76. In this replicon, the HCV IRES within the 5′ NTR drives translation of the first 15 amino acids of the core protein fused to neomycin (Neo). The EMCV IRES is 3′ of Neo and drives translation of an hRluc–ubiquitin (Ub)–foot-and-mouth disease virus 2a peptide (FMDV2a)–NS3–NS5B fusion protein (see Fig. 2A). The presence of FMDV2a and Ub increases the likelihood of cleavage of the hRluc reporter from the C-terminal HCV nonstructural proteins. Preparation of stable replicon cells was carried out as described for the genotype 2a replicon J6/JFHEMCVIRES2aRlucNeo.

FIG 2.

Clearance of HCV RNA from replicon cells by treatment with AP80978. (A) Schematic of the genotype 1b replicon [Con1/SG-Neo(I)hRluc2aUb] used in the replicon clearance assay. (B) Replicon cells were untreated or treated with AP80978 (15 μM or 3.75 μM) or control compounds (2.5 μM cyclosporine [CsA] and 7.5 μM VX950) in the absence of G418 selection for four passages. Cells were then cultured in selection media in the absence of compound to determine when replicon is present and to allow “rebound.” HCV RNA levels from each treatment and time point were quantified and are expressed as log₁₀ change compared with HCV RNA levels in untreated samples at day 0, with all values normalized relative to GAPDH RNA. No rebound was observed in replicon cells that had been cultured in the presence of CsA, VX950, or AP80978 during the clearance phase of the assay. Values are means ± standard deviations (error bars) from three replicates.

Con1/SG-Neo(I)hRLuc2aUb replicon cells were seeded into 12-well cluster plates at a density of 5 × 10⁴ cells/well in complete Huh-7.5 medium lacking G418 (Invitrogen). After 4-h attachment, triplicate wells were harvested for RNA extraction as a 0-h control. Additional wells were treated in triplicate with either 1% DMSO (untreated control), 7.5 μM VX950, 2.5 μM CsA, or 3.75 μM or 15 μM AP80978 (replicon clearance phase). Cell density was monitored daily, and when the cells reached 80% confluence, a sample was collected for RNA extraction, and a subset of the remaining cells were split at a 1:3 ratio into the appropriate compound-containing media. At the fourth passage, the cells were collected and passaged into media lacking compound in the presence of 250 μg/ml G418 to initiate the rebound phase. Untreated cells surviving the treatment were collected as they reached 80% confluence. RNA from all time points and treatments was extracted using a 96-well RNeasy RNA extraction kit (Qiagen, Valencia, CA). RNA levels of HCV 3′ NTR and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were quantified via quantitative reverse transcription-quantitative PCR (RT-qPCR). GAPDH RT-qPCR was carried out using the following primers: forward, 5′-CCTGCACCACCAACTGCTTA-3′; reverse, 5′-GCAGTGATGGCATGGACTGT-3′; probe, 5′-Cy5-CTGGCCAAGGTCATCCATGACAACT-BHQ-2-3′ (BHQ-2 stands for black hole quencher 2). HCV RNA levels were normalized to those of GAPDH and were expressed as HCV log₁₀ change relative to the HCV RNA level in untreated cells on day 0 using the comparative threshold cycle (CT) method (Applied Biosystems). At each time point, normalization of HCV to GAPDH was determined as follows: dCT = CT_HCV − CT_GAPDH. Calculation of ddCT value relative to t = 0 was determined as follows: ddCT = dCT_{time x} − dCT_{time zero}. HCV RNA log change was calculated as log(2^((−1) × ddCT)).

Evaluation of AP80978 against genotype 1a and 2a replicons.

The genotype 2a replicon, J6/JFHEMCVIRES2aRlucNeo, was created at Apath, LLC, and is encoded by the plasmid APP40. It is a stable subgenomic replicon with elements from both the JFH and J6 strains. The JFH IRES within the 5′ NTR drives translation of the first 19 amino acids of the JFH core protein linked to an hRluc-neomycin phosphotransferase (Neo) fusion protein. The EMCV IRES is 3′ of the hRluc-Neo ORF and drives translation of JFH NS3-NS5B, which is flanked at its 3′ end with the J6 3′ NTR. The differences between the JFH and J6 3′NTRs are within the variable region.

Preparation of stable genotype 2a replicon cells was performed as outlined above for CA32 cells except that 24 h after plating, the medium was removed and replaced with selection medium (i.e., complete medium containing 1 g/liter G418 (Invitrogen). Cells were monitored daily and passaged as needed to maintain a subconfluent culture. Selection was considered complete when cells electroporated in parallel with a polymerase-deficient replicon lacked any viable cells. After this observation, the concentration of G418 in the medium was decreased to 250 μg/ml for culture maintenance. Screening with the resulting genotype 2a replicon cells (APC140) was performed as described above for CA32 replicon cells except that the seeding density was 12,000 cells/well and the cells were assayed 48 h posttreatment in the absence of selection agent.

The genotype 1a replicon cells [Huh-7.5 cells containing the H/SG-Neo(L+I) replicon (APC89)] was previously reported (21). Replicon cells were seeded into 12-well plates at a density of 40,000 cells/well. After 4-h attachment, cells were treated in triplicate using a five-point threefold serial dilution series in 1% DMSO, which caused no adverse effects on cell viability for this assay duration in pilot studies. After 72 h, RNA was extracted using the RNeasy RNA extraction kit (Qiagen). The total RNA concentration was determined using Ribogreen RNA quantitation reagent (Invitrogen) and normalized between samples. HCV RNA levels were quantified via RT-qPCR using MultiCode-RTx HCV viral load primer mix and RNA reagent set (EraGen Biosciences, Madison, WI) on an Applied Biosystems 7300 real-time PCR system. EC₅₀s were determined as described above.

Evaluation of AP80978 against other viruses.

BHK-S cells harboring a subgenomic, puromycin-selectable Renilla luciferase reporter yellow fever virus (YFV) replicon (YF-hRUPac) (22) were seeded at 10,000 cells/well into 96-well plates. The serotype 2 dengue virus replicon was constructed by removing a structural protein-coding region between capsid gene codon 28 and the last 26 codons at the 3′ end of the envelope gene from an infectious clone of dengue 2 virus (DEN2) 16681 (23). This deletion was replaced with the humanized Renilla luciferase-ubiquitin-puromycin acetyltransferase (hRUPac) cassette from the yellow fever virus replicon (22) to generate pD2-hRUPac. BHK-S cells harboring the dengue virus replicon were seeded at 8,000 cells/well in 96-well plates. The respiratory syncytial virus (RSV) replicon is similar to that reported by Malykhina et al. (24) except that it also contains a Renilla luciferase reporter located between the green fluorescent protein (GFP) and NS1 coding sequence. Screening was carried out by seeding 1,500 RSV BHK-SR19–T7 replicon cells and 8,500 BHK-S parental cells/well in 96-well plates. For West Nile virus (WNV), BHK-WNV-Rep replicon cells (22) were seeded at 10,000 cells/well into 96-well plates. The Ebola virus (EBOV) minigenome expresses a Renilla luciferase reporter-encoding plasmid that carries cis-acting elements for Ebola virus replication. The EBOV minigenome was transfected into BHK-SINRep T7 cells, along with four expression plasmids that encode the replicase proteins NP, VP35, L, and VP30 (25). EBOV minigenome cells were seeded at 20,000 cells/well into 96-well plates.

BHK-S cells harboring a subgenomic, puromycin-selectable firefly luciferase reporter Sindbis virus replicon (SINrep19.FLuc.Pac) (26) were seeded at 10,000 cells/well into 96-well plates. The influenza A virus (FluA) minigenome expressing the firefly luciferase reporter carries cis-acting elements for FluA replication. The FluA minigenome was transfected into 293T cells, along with four expression plasmids that encode the replicase proteins NP, PA, PB1, and PB2. FluA minigenome cells (27) were seeded at 20,000 cells/well into 96-well plates.

For determination of antireplicon activity, all cells were treated with compounds using a five-point threefold serial dilution series with four replicate treatments per dose. A 24-h incubation period was chosen to maximize screening throughput, following which Renilla (dengue virus, YFV, RSV, WNV, or EBOV) or firefly (Sindbis virus or FluA) luciferase levels were evaluated to determine an EC₅₀. In parallel, cell toxicity/cell viability was assayed on parental cells using a CellTiter-Glo luminescent cell viability assay.

Cytotoxicity evaluation in other cell types.

HepG2 cells, Caco-2, MRC-5, and Jurkat cells were obtained from the American Type Culture Collection (ATCC) (Manassas, VA). HepG2 and MRC-5 cells were maintained in Eagle's minimal essential medium (Invitrogen) supplemented with 10% FBS, 100 units/ml penicillin, and 100 μg/ml streptomycin. Caco-2 cells were grown in the same medium but with 20% FBS. Jurkat cells were maintained in RPMI 1640 medium supplemented with 10% FBS, 100 units/ml penicillin, and 100 μg/ml streptomycin. All cells were incubated at 37°C in a humidified 5% CO₂ incubator. Cytotoxicity screening was carried out by incubating cells (20,000 cells/well for HepG2, Caco-2, and MRC-5 cell lines and 40,000 cells/well for Jurkat cells) in 96-well plates in the presence of serial dilutions of compound for 24 h, a time point chosen as a realistic interval to observe control inhibitor activity and which also maximized assay throughput. At the end of the incubation period, cell viability was assessed by the CellTiter-Glo luminescent cell viability assay.

Generation and evaluation of AP80978-resistant replicons.

Clone A cells (28) contain a genotype 1b (Con1) stable subgenomic replicon, and these cells were used for resistant mutant selection. The cells were seeded at a density of 1 × 10⁶ cells per 75-cm² T-flask. After 4-h attachment, cells were treated with complete medium supplemented with either 1% DMSO (untreated control) or 10 μM AP80978. The medium was changed every 3 or 4 days to replenish compound and G418, and the cells were split as needed to maintain subconfluent cultures. After 12 days of culture in the presence of 10 μM AP80978, the concentration was increased to 20 μM, and cells were cultured for an additional 19 days. After this period, a sample of cells from both AP80978-treated and control cells were collected for evaluation of their response to AP80978 as follows. The cells were seeded into 12-well plates at a density of 40,000 cells/well in the absence of G418. After 4-h attachment, the cells were treated in triplicate with AP80978 using a five-point, threefold serial dilution series in 1% DMSO. After 72 h, total RNA was extracted using the RNeasy RNA extraction kit (Qiagen). Total RNA was quantified using the Ribogreen RNA quantitation reagent (Invitrogen), and concentrations were normalized between samples. HCV RNA levels were quantified via quantitative RT-PCR using MultiCode-RTx HCV viral load primer mix and RNA reagent set on an Applied Biosystems 7300 real-time PCR system. EC₅₀s were determined as described above.

Evaluation of cell-associated versus replicon-associated resistance.

Total RNA (10 μg) extracted from control and AP80978-treated cells was electroporated into the clone A parental cell line Huh7. Cells were selected with 1 mg/ml G418, and after selection was complete (determined by complete cell death in cells electroporated without RNA), the cells were pooled and evaluated for their response to AP80978 as described above.

Replicon sequence analysis.

Total RNA was extracted from control and AP80978-resistant replicon cell lines using the RNeasy RNA extraction kit (Qiagen). RT-PCR was carried out with replicon-specific primers using SuperScript III reverse transcriptase (Invitrogen) and iProof high-fidelity polymerase (Bio-Rad). PCR amplicons were sequenced and compared against the wild-type reference sequence using Clone Manager software. The four mutations identified within NS4B in the AP80978-resistant cells that encode G60 [no change or silent], F98V [F at position 98 changed to V], F98L, and S238Y were introduced individually into the genotype 1b replicon-encoding plasmid APP76 using standard PCR-based techniques. The mutated plasmids were sequenced to verify that only the desired mutations had been introduced.

Colony formation assay.

IVT RNA corresponding to wild-type replicons or replicons with F98V or F98L resistance mutations were electroporated into Huh-7.5 cells as described for the genotype 2a replicon. Electroporated cells (225, 450, 900, or 1,800) were seeded in triplicate into 6-well cluster plates along with cells electroporated with a polymerase-defective construct to achieve a total cell number of 90,000 cells/well. After 24-h attachment, the cells were treated with either 1 mg/ml G418 alone or G418 in the presence of 20 μM AP80978. The medium was changed every 3 or 4 days to replenish compound and remove dead cells. After 2 to 3 weeks, the media were removed, the cells were fixed with 7% formaldehyde and stained with 1% crystal violet (Sigma-Aldrich) in 50% ethanol, and the colonies were counted.

Generation and testing of intergenotypic 2a/1b replicons.

Standard PCR-based techniques were used to produce a series of genotype 2a intergenotypic replicons (encoded by plasmid APP40 [described above]) in which the genotype 2a NS4B-encoded amino acids were replaced with the corresponding sequence from genotype 1b: DNA fragment from the genotype 2a replicon encoding genotype 1b NS4B amino acids 7 to 254 [NS4B(Con1_7-254)], NS4B(Con1_7-52), NS4B(Con1_53-254), NS4B(Con1_219-254), NS4B(Con1_{7-52, 219-254}), and NS4B(Con1_53-219). Polymerase-defective versions of each construct were prepared in parallel to evaluate replication of each intergenotypic replicon. The resulting plasmids were sequenced to verify that the desired constructs had been produced. Preparation of intergenotypic replicon cells and compound screening were performed as outlined above for genotype 2a replicon cells. The genotype 2a replicon containing the NS4B configuration NS4B(Con1_7-254) did not replicate by comparison with the polymerase-defective counterpart.

Generation and testing of AP80978-sensitive virus.

To construct an AP80978-sensitive hepatitis C virus, a 3,355-bp SpeI-RsrII (NEB) DNA fragment from the genotype 2a replicon, encoding genotype 1b NS4B amino acids 53 through 219 [i.e., NS4B(Con1_53-219)] was excised and inserted in the same restriction sites in the genotype 2a/2a J6/JFH Jc1 (29) HCV plasmid construct (see Fig. 6A). This HCVcc construct is referred to as J6/JFH Jc1/Con 1 NS4B_53-218.

FIG 6.

Effects of AP80978 and control compounds on virus replication. (A) Schematic of the AP80978-sensitive virus J6/JFH Jc1 (Con1 NS4B_53-218). (B) Huh-7.5 cells were infected with J6/JFH Jc1 in the presence of the indicated concentrations of AP80978 or its structural analogs. After 48 h, HCV RNA levels from each treatment were quantified by RT-qPCR and are expressed as log₁₀ change compared with HCV RNA levels in untreated samples. (C) Huh-7.5 cells were infected with J6/JFH Jc1 in the presence of the indicated concentrations of the control compound 2′C-meA, VX950, or cyclosporine (CsA). Experimental procedure and data analysis are as described above for panel B. (D) Huh-7.5 cells were infected with J6/JFH Jc1 (Con1 NS4B_53-218) in the presence of the indicated concentrations of AP80978 or its structural analogs. Experimental procedure and data analysis are as described above for panel B. (E) Huh-7.5 cells were infected with J6/JFH Jc1 (Con1 NS4B_53-218) in the presence of the indicated concentrations of control compound 2′C-meA, VX950, or CsA. Experimental procedure and data analysis are as described above for panel B.

The resulting plasmid was linearized with XbaI (NEB), and DNA was prepared for in vitro transcription and electroporation as described above for the replicons. Following electroporation, cell culture supernatants were collected and concentrated using Amicon centrifugal filter units with a 100,000-molecular-weight cutoff (EMD Millipore Corporation, Billerica, MA). The virus titer was determined by infecting replicate wells containing Huh-7.5 cells with serially diluted virus, staining 4 days later with NS5A monoclonal antibody, and calculating the 50% tissue culture infective dose (TCID₅₀)/ml from the number of positively stained replicates for each dilution.

To evaluate the effect of AP80978 on J6/JFH Jc1/Con1 NS4B_53-218 virus, Huh-7.5 cells were seeded into 96-well plates at a density of 20,000 cells/well. After 24-h attachment, cells were treated in quadruplicate with serially diluted compound in 1% DMSO in the presence of 2,000 TCID₅₀ units/well virus (multiplicity of infection [MOI] of 0.1). Following 48 h of incubation, the cells were washed and RNA was extracted using RNeasy RNA extraction kit (Qiagen). RNA levels of HCV 3′ NTR were quantified via RT-qPCR and were expressed as HCV log change relative to untreated control cells using the comparative threshold cycle method.

RESULTS

Identification of AP89652 in a genotype 1b replicon assay.

In order to identify new classes of antiviral compounds that inhibit HCV replication, a high-throughput cell-based screen was carried out using a transient genotype 1b subgenomic Renilla luciferase replicon (Fig. 1A). A primary screen of 93,000 commercially available compounds against these replicon cells identified 209 compounds that decreased Renilla luciferase levels by 50%. As the primary screen did not employ a method to identify cytotoxic compounds, these 209 compounds were further evaluated in a secondary assay to calculate efficacy (EC₅₀) and toxicity (CC₅₀). Evaluation of these compounds in the secondary screen identified 38 compounds with EC₅₀s less than 10 μM and CC₅₀s greater than 100 μM. The 38 compounds belonged to 13 distinct chemical classes, with 15 compounds categorized as orphan compounds. One compound, AP89652 (Fig. 1B), that belongs to a tetrahydropyrazole-[1,5,a]-pyrimidine chemical series was selected for further study based on its favorable physicochemical properties, novelty, and amenability of this chemical series to SAR (structure-activity relationship) studies.

AP89652 contains two chiral centers (Fig. 1B), yielding four potential isomers. However, based on the literature describing the synthesis of tetrahydropyrazolo-[1,5a]-pyrimidine chemical series (15) and our results, it was found that AP89652 is a racemic mixture of two syn enantiomers, as shown in Fig. 1C. To determine which enantiomer of AP89652 was responsible for HCV replication inhibitory activity, the optically pure enantiomers were isolated via chiral chromatography and evaluated against the replicon cell line. Screening against CA32 genotype 1b subgenomic replicon cells revealed that only one of the two enantiomers, AP80978, had replicon-inhibitory activity, with an EC₅₀ of 0.63 μM, while the other, AP80977, lacked activity (Table 1). Similar results were obtained when replicon levels were quantified using RT-qPCR (data not shown).

TABLE 1.

Activity of AP89652 isomers in replicon assays^a

Compound	Description	EC50 (μM)	CC50 (μM)	SI
AP89652	Racemate	1.78	>100	>56
AP80977	Enantiomer 1	>25	>100	NA
AP80978	Enantiomer 2	0.63	>100	>158

^aEC₅₀, the concentration at which a 50% reduction in Renilla luciferase levels was observed relative to an untreated control; CC₅₀, the concentration resulting in a 50% decrease in cell viability; SI, selectivity index; NA, not applicable.

HCV replicon clearance by AP80978.

The HCV replicon clearance assay served as a cell culture model for virus clearance. At the onset of the assay, cells harboring a stable HCV replicon were cultured under nonselection conditions in the presence or absence of compound. During this “replicon clearance phase,” cells were able to proliferate independently of HCV replicon replication. The “rebound phase” was initiated by inhibitor removal and resumption of G418 selection, allowing growth of only HCV replicon-harboring survivor cells.

In this assay, the cells harboring the stable genotype 1b subgenomic replicon Con1/SG-Neo(I)hRluc2aUb (Fig. 2A) were treated with either 15 μM or 3.75 μM AP80978 in the absence of neomycin for 17 days. To compare the efficacy of AP80978 with those of other antiviral compounds, parallel cultures were treated with 2.5 μM CsA or 7.5 μM VX950. Compound-treated cells were passaged as needed to maintain a subconfluent culture, replenishing compound and collecting an aliquot of cells at each passage. On day 17, cells were seeded in G418-containing selection media in the absence of compound, and cells were cultured for an additional 7 days, during which untreated cells approached confluence. Massive cell death was observed in cultures that had been treated with compounds. At each passage, HCV RNA levels were quantified and normalized relative to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) via RT-qPCR for treated and control samples. HCV RNA levels remained similar over the course of the assay in control cells grown in the absence of inhibitors, while cells treated with both doses of AP80978 exhibited a progressive decrease in HCV RNA levels during the replicon clearance phase, with a reduction of 1.5 to 1.75 log₁₀ RNA copies by day 17 of the assay (Fig. 2B). Similar to results with AP80978, treatment with CsA and VX950 was associated with a maximal decrease in HCV RNA levels of 1.5 to 2 log₁₀ units. During the rebound phase, HCV RNA levels in all inhibitor-treated cells, including AP80978 and control CsA and VX950, were undetectable in the presence of G418, indicating the absence of functional replicon RNA within these cells (Fig. 2B). In control untreated cells, replicon RNA was slightly lower than baseline but within the variation typically observed in RT-qPCR in the clearance assay.

Genotype and virus specificity of AP80978.

To evaluate the genotype specificity of AP80978, the compound was screened against cells harboring the genotype 1a replicon H/SG-Neo(L+I) (21) and the genotype 2a replicon J6/JFHEMCVIRES2aRlucNeo. AP80978 inhibited replication of the genotype 1a replicon, with an EC₅₀ (0.62 μM) similar to that for the genotype 1b replicon (1.76 μM). However, the compound was completely inactive against the genotype 2a replicon (EC₅₀ of >25 μM). Screening against genotype 1a, 1b, and 2a replicons by two external laboratories confirmed similar EC₅₀s for genotypes 1a and 1b and inactivity against genotype 2a (data not shown).

To evaluate whether AP80978 was active against other viruses, the compound was tested against virus-free minigenome and replicon systems that have been developed for high-throughput screening to identify replication inhibitors of negative-strand viruses and other positive-strand viruses. AP80978 was inactive against infection-free negative-strand RNA virus systems for Ebola virus, influenza A virus, and respiratory syncytial virus, although the controls (ribavirin, CsA, and 2′C-meA) for each of these systems behaved as expected (Table 2). Similarly, the compound was inactive against the alphavirus Sindbis virus replicon, as well as other positive-strand flavivirus replicons for dengue virus, yellow fever virus, and West Nile virus, indicating that AP80978 is an HCV-specific antiviral compound. To confirm that the compound that was used in these studies was active, AP80978 was tested in parallel against HCV replicon cells, and it showed efficacy.

TABLE 2.

Activity of AP80978 against other viruses^a

Compound	Replicon EC50 (μM)					Minigenome EC50 (μM)
	Dengue virus	YFV	Sindbis virus	RSV	WNV	Ebola virus	FluA
AP80978	>20	>20	>20	>20	>20	>20	>20
Ribavirin	2.0	3.7	2.1	3.7	ND	10.1	20.0
CsA	12.0	8.6	>20	2.7	5.5	5.4	10.2
2′C-meA	4.5	2.9	5.3	>20	4.0	>20	>20

^aAbbreviations: YFV, yellow fever virus; RSV, respiratory syncytial virus; WNV, West Nile virus; FluA, influenza A virus; CsA, cyclosporine; ND, not determined.

Cytotoxicity of AP80978 in different cell types.

To determine the effect of AP80978 on other cell types, cytotoxicity was evaluated against a panel of cell lines from an array of different tissues. Similar to observations using the Huh-7.5 cell line that supports HCV replication in cell culture, AP80978 was nontoxic against another human liver cell line, HepG2, with a CC₅₀ value of >100 μM. It was also nontoxic (had CC₅₀ values of >100 μM) in human cell lines derived from intestinal epithelium (colorectal) (Caco-2), lung fibroblasts (MRC-5), and T lymphocytes (Jurkat).

Mechanism of action of AP80978. (i) Selection and characterization of AP80978 resistance mutations.

Preliminary biochemical studies showed that AP80978 had no activity against the HCV NS3-4A serine protease and NS5B polymerase (not shown), suggesting a novel viral or host target. Due to the error-prone nature of the HCV RNA polymerase, culture of HCV replicon cells in the presence of replication inhibitors can result in the selection of inhibitor-resistant replicons. In an effort to gain insight into the molecular target of AP80978, clone A cells resistant to AP80978 were selected by sequential passaging in the presence of compound. Clone A cells are a human hepatoma cell line that contains a stable genotype 1b (Con1 strain) subgenomic replicon (28). Cells were cultured under neomycin-containing conditions for 12 days in medium supplemented with 10 μM AP80978, followed by culture for an additional 19 days in 20 μM AP80978. Control clone A cells grown in the presence of diluent only were cultured in parallel. After this culture period, cells were evaluated for their response to AP80978 by culturing treated and control cells in the presence of serial dilutions of AP80978 for 72 h and subsequently quantifying HCV RNA levels via RT-qPCR to determine the EC₅₀. Clone A cells that had undergone selection exhibited >10-fold decreased sensitivity to AP80978 (EC₅₀ > 20 μM) relative to the control cells (EC₅₀ = 1.8 μM).

To determine whether resistance was encoded by the replicon, total RNA was extracted from both AP80978-resistant and control cells, and RNA was reintroduced via electroporation into naive Huh7 cells, the parental cell line of clone A cells. Electroporated cells that harbored actively replicating replicons were selected with G418, pooled, and assayed for sensitivity to AP80978 as described above. Cells electroporated with RNA from control cells were sensitive to AP80978 (EC₅₀ = 0.91 μM), while those electroporated with RNA from AP80978-resistant cells maintained resistance to the compound (EC₅₀ > 20 μM), indicating that the resistance was associated with the replicon.

To identify nucleotide changes that conferred resistance to AP80978, RNA extracted from AP80978-resistant and control clone A cells was reverse transcribed and amplified using replicon-specific primers. Sequence analysis of the amplified replicons revealed four point mutations present within the sequence encoding NS4B from the AP80978-resistant cells that were absent from the control or wild-type replicon sequence. Three of the four mutations resulted in amino acid changes corresponding to F98L, F98V, and S238Y within NS4B. One silent mutation at codon 60 was also observed.

Each mutation was introduced independently into a stable genotype 1b subgenomic reporter replicon plasmid (Fig. 3A) via site-directed mutagenesis. Sequencing of the resulting plasmids revealed that they were void of additional mutations. IVT RNA generated from each mutant and parental replicon plasmid was introduced into Huh-7.5 cells, a highly permissive human hepatoma cell line. The resulting replicon cell lines were evaluated for response to AP80978 and the positive-control compounds CsA and VX950. A differential response was observed between cells electroporated with different replicon RNAs. Cells containing replicons with F98V and F98L amino acid substitutions were resistant to AP80978 (EC₅₀ > 20 μM), while those containing parental wild-type replicon (EC₅₀ = 0.29 μM), the S238Y substitution (EC₅₀ = 0.41 μM), or the silent mutation (G60nc) (EC₅₀ = 0.24 μM) remained sensitive to AP80978 (Table 3). These data suggest that NS4B is the target of AP80978 and that changing the amino acid at residue 98 of NS4B to either valine or leucine was sufficient to confer resistance to the compound. Cells electroporated with IVT RNA from each construct were responsive to CsA (EC₅₀s ranged from 0.1 to 0.16 μM) and VX950 (replicons cleared from all cell lines [data not shown]), indicating that resistance to AP80978 did not confer cross-resistance to either of these compounds.

FIG 3.

Localization of AP80978-selected amino acid substitutions within NS4B. (A) Schematic of the genotype 1b replicon present within clone A cells that was used to select AP80978-resistant replicons. The locations of amino acid substitutions within NS4B (G60 no change [G60nc], F98V, F98L, and S238Y) are indicated. (B) Alignment of NS4B from JFH1 (genotype 2a) and Con1 (genotype 1b), revealing the presence of a leucine residue at amino acid 98 in JFH1 (boxed). Amino acids in JFH1 that are identical to those in Con1 are indicated by dashes. A mutation to encode leucine at this position in genotype 1b replicons confers resistance to AP80978.

TABLE 3.

Activity of AP80978 against genotype 1b replicons containing amino acid substitutions

Replicon	EC50 (μM)
Wild-type	0.29
G60nca	0.24
F98V	>20
F98L	>20
S238Y	0.41

^aG60nc indicates a silent mutation at the G60 codon.

Alignment of the NS4B amino acid sequences from genotypes 1b and 2a revealed a leucine at codon 98 of NS4B in genotype 2a (Fig. 3B). Since mutation that changed phenylalanine to leucine at this position in the genotype 1b replicon conferred resistance to AP80978, mutagenesis was carried out to determine whether a phenylalanine at this position in the genotype 2a replicon was sufficient to confer sensitivity to the compound. The resulting genotype 2a replicon with the L98F substitution remained resistant to the compound (EC₅₀ > 20 μM), indicating that a mutation resulting in amino acid substitution in this residue alone was insufficient to confer AP80978 sensitivity to an HCV genotype 2a subgenomic replicon.

The effects of AP80978 resistance mutations were evaluated in a long-term colony formation assay. IVT RNA from either wild-type or resistant genotype 1b replicon constructs was electroporated into naive Huh-7.5 cells and plated at various densities in the presence of cells that had been electroporated with polymerase-defective constructs in order to allow transient growth of cells in the presence of G418. Cells that had been electroporated with all constructs grew in the presence of G418, with similar numbers of colonies. However, only cells that had been electroporated with IVT RNA encoding resistance mutations grew in the presence of AP80978 (Fig. 4). In cells electroporated with RNA encoding the F98L resistance-associated amino acid change, there were fewer and smaller colonies in the presence of AP80978, suggesting that maintaining resistance to this compound may be associated with a fitness cost. However, colony number and size were similar in the presence and absence of AP80978 for cells electroporated with the construct encoding the F98V substitution.

FIG 4.

Colony formation of wild-type and AP80978-resistant replicons in the presence of AP80978. (Top) Images of a colony formation assay in which cells were electroporated with wild-type or AP80978-resistant (F98V or F98L) replicon RNA, plated in triplicate (1,800 cells/well) in the absence (G418 only) or presence (G418 + AP80978) of AP80978 under selection conditions, and surviving colonies were stained after 3 weeks. (Bottom) Graphic representations of colony formation assays in which 225, 450, 900, and 1,800 electroporated cells were plated and cultured in the absence (−AP80978) or presence (+AP80978) of AP80978. Values are means ± standard deviations (error bars) from three replicates.

(ii) Evaluation of AP80978 against intergenotypic genotype 2a/1b replicons.

To confirm NS4B as the molecular target of AP80978, an alternate approach was taken that took advantage of the differential response of genotypes 1b and 2a to AP80978. Using the plasmid encoding J6/JFHEMCVIRES2aRlucNeo (a genotype 2a replicon that is insensitive to AP80978) as a template, six intergenotypic constructs were made, in which regions encoding the following amino acids of NS4B were replaced with the corresponding amino acids from genotype 1b NS4B (sensitive to AP80978): (i) 7 to 254, (ii) 7 to 52, (iii) 53 to 254, (iv) 219 to 254, (v) 7 to 52 and 219 to 254, and (vi) 53 to 218 (Fig. 5). Testing the resulting intergenotypic constructs would show which portions of genotype 1b NS4B protein confer AP80978 response. IVT RNA generated from each of the six intergenotypic replicon cDNAs as well as the genotype 2a parental replicon was electroporated into Huh-7.5 cells and evaluated for response to AP80978 and the control compound CsA (Fig. 5). The intergenotypic replicon that contained the nearly full-length NS4B from genotype 1b (amino acids 7 to 254) did not replicate, similar to the replication-defective control replicon, while the remaining intergenotypic replicons replicated to various degrees based on reporter levels. Compound screening was initially carried out in unselected cells 96 h postelectroporation, and again following selection with G418, with similar results. Two intergenotypic replicons that encoded amino acid sequences from genotype 1b within the central region of NS4B (amino acids 53 to 254 and amino acids 53 to 218) were sensitive to AP80978, with EC₅₀s less than 1 μM, while all other intergenotypic replicons, as well as the genotype 2a parental replicon, did not respond to AP80978 treatment. Thus, the molecular target of AP80978 was located between amino acids 53 and 218 within NS4B of genotype 1b. All of the intergenotypic and wild-type control constructs were responsive to CsA (not shown).

FIG 5.

Screening AP80978 against genotype 2a (GT2a) replicons containing intergenotypic GT2a/1b NS4B. Schematics of intergenotypic NS4B regions within the genotype 2a replicon are illustrated, with black boxes representing genotype 1b regions and gray boxes representing genotype 2a regions. The AP80978 EC₅₀s are indicated to the right of each replicon construct. The replicon containing Con1 amino acids 7 through 254 (i.e., NS4B Con1_7-254) failed to produce cells that survived G418 selection.

(iii) Evaluation of AP80978 activity against cell culture-produced HCV.

To evaluate the effect of AP80978 on virus replication in the context of the fully infectious replication cycle, an AP80978-sensitive virus construct was produced by cloning residues 53 to 218 from genotype 1b into a J6/JFH1 Jc1 construct, to produce J6/JFH Jc1 (Con1 NS4B_53-218) (Fig. 6A). IVT RNA produced from this construct was electroporated into Huh-7.5 cells, and intergenotypic virus was collected from the cell culture medium. Titers were low, suggesting that the virus replicated poorly. Huh-7.5 cells were infected with either the intergenotypic or wild-type virus in the presence of either AP80978, AP80977 (the second syn enantiomer that was inactive in the replicon assay), AP89652 (the racemic mixture), or AP80935 (an analog of AP80978 with an aromatic six-membered ring). Parallel cultures were treated with the HCV inhibitors 2′C-meA, VX950, and CsA. After 48 h, RNA was extracted, and HCV RNA levels were quantified. As expected, in cells infected with the J6/JFH Jc1 virus, HCV RNA levels did not change appreciably in cells treated with 0.06 to 5 μM AP80978, AP80977, AP89652, or AP80935 (Fig. 6B), although high concentrations of the control inhibitors 2′C-meA and VX950 were associated with a 2 to 3 log₁₀ unit reduction (Fig. 6C). Lower inhibitory potency was observed for CsA for J6/JFH Jc1 HCV RNA (Fig. 6C). In cells infected with intergenotypic J6/JFH Jc1 (NS4B_53-218), HCV RNA levels decreased >1 log₁₀ unit at 70 nM AP80978 and ∼0.7 log₁₀ unit at 20 nM (Fig. 6D). Unexpectedly, AP80977, the enantiomer that was inactive in the genotype 1b replicon assay, had inhibitory activity against the intergenotypic virus, with a reduction in HCV RNA of ∼0.7 log₁₀ unit at 0.67 μM. Treatment with AP89652 and AP80935 also resulted in decreases in HCV RNA levels, but their potencies were less than that of AP80978. However, the potency of AP80935 was similar to that of AP80978 in a 24-h genotype 1b replicon assay (not shown). One key finding is that AP80978 maximally decreased HCV RNA by 1.5 log₁₀ unit at 220 nM, which was lower than the replicon assay EC₅₀ and suggestive of enhanced activity in the context of replicating virus. At a comparable concentration (190 nM), inhibitory efficacy was decreasing or lost for the HCV inhibitors 2′C-meA, VX950, and CsA (Fig. 6E).

DISCUSSION

The purpose of this study was to identify new classes of anti-HCV compounds, particularly ones with novel viral targets. Such compounds may be critical in patients failing existing therapies and/or patients who develop drug-resistant viruses. We identified a novel small-molecule inhibitor, AP80978, with activity against HCV genotype 1a and 1b, but not genotype 2a, which was not toxic in multiple cell lines. AP80978 was similar to VX950, an approved therapeutic for chronic HCV infection, in potency and clearance rate in the replicon clearance assay and against replicating virus. Preliminary studies indicated that AP80978 activity was not directed against HCV protease and polymerase, suggesting a mechanism of action distinct from many compounds currently being evaluated in clinical studies. Through two complementary approaches, generating resistant mutants and evaluating the differential response of intergenotypic genomes to the compound, we demonstrated that AP80978 acts against a novel target, NS4B.

Experimental data from numerous studies suggest several molecular functions for the NS4B protein, including (i) formation of a membranous web, the modified membrane structure located in the endoplasmic reticulum that is the proposed site for HCV replication (30), (ii) modulation of NS5B RNA-dependent RNA polymerase activity (31), (iii) modulation of HCV and host cellular translation (32–34), (iv) nucleotide binding and GTPase activity (35), (v) modulation of NS5A hyperphosphorylation (36, 37), and (vi) HCV RNA binding (38). In theory, antiviral compounds could target any of these functions of NS4B.

Other labs have recently identified compounds that inhibit HCV replication by targeting NS4B, attesting to its potential as an effective drug target against HCV (reviewed in reference 39). Clemizole was identified after screening a compound library for inhibitors of NS4B-HCV RNA binding by using in vitro protein expression coupled with a microfluidic affinity analysis (38). This compound inhibited HCV replication in cell culture, and mutations conferring resistance to this compound displayed increased affinity for the viral RNA. Viropharma discovered several classes of compounds that interacted with NS4B using a biochemical binding assay and demonstrated antireplicon activity, presumably by apoptosis induction of NS4B-expressing cells (40). A new class of NS4B inhibitors was identified at Stanford University by screening for inhibitors of vesicle aggregation, which is mediated by the NS4B AH2 domain, an amphipathic helix between amino acids 42 and 66 that is involved in membrane rearrangement (41). Of particular interest is anguizole, an NS4B antagonist that targets the AH2 domain and changes its subcellular distribution (42). Anguizole and AP80978 share two-dimensional structural similarity in their pyrazolopyrimidine core and substitution nature and pattern, and it would be interesting to determine whether they have similar mechanistic activity. A novel NS4B inhibitor with activity against genotype 1a and 1b replicons is PTC725, a small molecule whose inhibitory activity is additive to synergistic with alpha interferon and HCV protease and polymerase inhibitors, although activity against replicating HCV was not assessed (43). Similar to our data, F98 was implicated in the response to PTC725, with mutants with F98L/C amino acid substitutions reduced in both inhibitor susceptibility and replicon fitness compared to the wild type (43). In that study, immunofluorescence experiments of replicon cells showed no changes in intracellular distribution of NS4B in the presence of the inhibitor or with replicons expressing F98C-substituted NS4B (43). As AP80978 and PTC725 are dissimilar in chemical structure, it would be informative to compare their three-dimensional models to determine whether they share similar electrostatic, van der Waals, and/or hydrophobic interactions with the target, since both compounds select for mutations encoding resistance at the same amino acid position. In this context, it will be worthwhile to further understand the mechanistic properties of AP80978, particularly whether it antagonizes NS4B by affecting interactions with HCV RNA and/or membrane rearrangements. Finally, the fact that AP80978-resistant replicons remained sensitive to CsA and VX950 suggests that drug combinations may be very effective. In this regard, it would also be important to know whether AP80978 is potentially additive or synergistic when combined with other inhibitors.

One unexpected finding from our study was the increased potency of both AP80978 and AP80977 in the virus assay relative to the replicon assay. In the replicon assay, AP80977 was inactive, while in the infection assay using cell culture-produced intergenotypic J6/JFH Jc1 (Con 1 NS4B_53-218) virus, its activity was similar to that of the control HCV inhibitors. Similarly, AP80978, which was already active in the replicon assay, had enhanced potency in the infection assay. The greater inhibitory effect on infectious virus versus replicon is unsurprising, because in addition to creating RNA genome replication foci, NS4B also interacts with other replicase proteins as well as itself during infectious virion production (44–46). Therefore, an inhibitor targeting NS4B may block numerous steps in the context of fully infectious virus replication. Whether this assay or replicon-based assays more closely reflect inhibitor performance in patients with chronic HCV infection is unknown. It is also not known whether the lower potency of the racemate in the replicon assay was due to dilution of the activity of AP80978 by AP80977 or because of an inhibitory effect of AP80977 on the activity of AP80978. Overall, however, the potency of AP80978 against fully replicating HCV does provide promise for in vivo efficacy and future mechanistic studies.

Additional studies to determine how this compound affects NS4B function, as well as structure-activity relationship studies to further improve potency and genotype coverage are warranted. Since genotype 1 is the most prominent HCV in North America and Europe and it is also the most difficult to treat, AP80978 provides a scaffold for development of newer inhibitors with pan-genotypic activity. Additionally, we believe that AP80978 will be a useful tool to help dissect the multiple roles of NS4B in the HCV replication cycle and aid the design of efficacious antiviral therapeutics targeting nonenzymatic viral replicase components.