Cell culture.

The reporter cell line Huh-luc/neo-ET (termed ET) (34) was provided by Ralf Bartenschlager (University of Heidelberg, Germany). This cell line harbors the persistently replicating I389luc-ubi-neo/NS3-3′/ET replicon containing the firefly luciferase gene-ubiquitin-neomycin phosphotransferase fusion protein and encephalomyocarditis virus (EMCV) internal ribosomal entry site (IRES)-driven NS3-5B HCV coding sequences containing the ET tissue culture-adaptive substitutions (E1202G, T1208I, and K1846T). The cell line was cultured as described previously (34).

HCV antiviral and cytotoxicity assays.

For in vitro antiviral and cytotoxicity assays, HCV replicon cells were seeded into 96-well tissue culture plates at 5.0 × 10³ cells per well in a volume of 85 μl per well and incubated at 37°C in 5% CO₂ for 24 h. Following 24 h of incubation, serial dilutions of miravirsen or a negative-control oligonucleotide (SPC4729) were added directly to the cells in triplicate wells (gymnosis [35]). IFN-α2b was included as a positive control. Following 48 h of incubation, the plates were assessed for anti-HCV activity by measurement of luciferase reporter activity (Britelite plus luminescence reporter gene kit [PerkinElmer, Shelton, CT]) with a Wallac 1450 Microbeta Trilux liquid scintillation counter. Percent inhibition for each drug concentration relative to the untreated replicon cells was calculated from the mean relative light units (RLU) of the triplicate wells containing a given drug concentration as a percentage of the mean RLU from six replicate wells containing untreated replicon cells. The 50% and 90% effective concentrations (EC₅₀ and EC₉₀) were calculated by linear regression. Data are expressed as the mean EC₅₀ ± standard deviations together with the range representing the low to high experimental values. In parallel, separate plates were assessed for cytotoxicity by XTT dye reduction. The 50% cytotoxic concentration (CC₅₀) was calculated as the concentration of drug that decreased the percentage of formazan produced by infected, drug-treated replicon cells to 50% of that produced by drug-free replicon cells.

In vitro combination studies.

For in vitro combination studies, HCV replicon cells were cultured as described above. Antiviral activity (EC₅₀) and cytotoxicity (CC₅₀) were determined for each of the compounds alone or in combination in triplicate wells on 3 separate plates. Following the 24-h incubation, eight successive 2-fold dilutions of miravirsen in cell culture medium without G418 were evaluated alone or in all possible combinations with five successive 2- or 5-fold serial dilutions of a second compound. The second compound included BMS-790052 (NS5A protein inhibitor), VX-222 (nonnucleoside NS5B inhibitor), telaprevir (NS3 protease inhibitor), BILN-2061(NS3 protease inhibitor), and 2′-MeC (nucleoside NS5B inhibitor). Concentrations of each drug were selected that bracketed the EC₅₀, with high concentrations being 1.2 to 2.4 μM, 10 U/ml, 10 nM, 148 μM, 32 pM, 4.5 μM, and 1.0 μM for miravirsen, IFN-α2b, VX-222, RBV, BMS-790052, 2′-MeC, and telaprevir, respectively. Six wells in each plate also received medium alone as a no-treatment control. Following 48 h of incubation, the plates were assessed for anti-HCV activity and cellular cytotoxicity as described above. Data from each experiment were analyzed by the technique of Prichard and Shipman, Jr. (36), using MacSynergy II, version 1.0 (36, 37). The difference between the observed antiviral effects from each combination and those expected if the interactions occurred independently (additive) are expressed as a volume (μM²%) above or below a plane that represents no interactive (additive) effects. Volumes with positive values at the 95% confidence interval (P < 0.05) indicate synergy, while volumes with negative values indicate antagonism. As a general guideline (37), volumes of between 50 and 100 μM²% suggest moderate synergy or antagonism, and volumes of >100 μM²% suggest strong synergy or antagonism.

In vitro antiviral activity against drug-resistant HCV replicons.

The in vitro antiviral activity of miravirsen was evaluated against HCV genotype 1b reporter replicons engineered to contain key amino acid substitutions in NS3 protease (A156T and R155K), NS5B polymerase (S282T and M423I), and NS5A protein (Y93H). Substitutions were constructed by site-directed mutagenesis of a bicistronic HCV replicon containing both a poliovirus IRES sequence at the 5′ end to increase luciferase translation and RNA replication and cell culture-adaptive substitutions in NS3 (E1202G and T1280I) and NS4B (K1846T) (38). In vitro-transcribed RNA from the mutated plasmids was used in transient-transfection assays of Huh 7 cells. Five reference compounds (BMS-790052 [NS5A], VX-222 [NS5B], telaprevir [NS3], BILN-2061[NS3], and 2′-MeC [NS5B]) also were included. Briefly, Huh7 cells were transfected with either the wild-type or the mutant RNA constructs by electroporation and subsequently seeded into 96-well black tissue culture plates at a cell culture density of 2.0 × 10⁴ cells per well in a volume of 100 μl. Reference compounds and miravirsen were tested in 9-point, half-log (3.2-fold) serial dilutions. Luciferase activity was measured 72 h after compound treatment, and EC₅₀s were determined from dose-response curves. Fold resistance was expressed as the ratio of the EC₅₀ for mutant HCV replicon to the EC₅₀ for the wild-type HCV replicon.

In vitro selection studies.

The HCV reporter cell line Huh-luc/neo-ET was serially passaged in the presence of 750 μg/ml G418 alone or G418 with fixed and escalating concentrations of miravirsen, SPC4729, or telaprevir (positive control) for time periods of up to 148 days (see Fig. 3). Concentrations ranged from 1.00 to 80 μM (2- to 160-fold the mean miravirsen EC₅₀ based on three independent determinations), 1.00 to 80 μM, and 0.60 to 6 μM (2- to 20-fold the telaprevir EC₅₀) for miravirsen, SPC4729, and telaprevir, respectively. Antiviral assays on passaged cultures were performed as described above using the luciferase assay or by quantitative reverse transcription-PCR (RT-PCR). Sequence analysis was performed on stored samples from G418 control-, miravirsen-, or SPC4729-passaged cells at days 39, 72, 117, 128, and 148 of the serial passage study. Amplification and sequence analysis of miR-122 binding sites (S1, S2, and S3) in the HCV 5′- and 3′UTRs was performed initially by site-specific primed RT-PCR followed by population-based sequencing. Briefly, total RNA was extracted and purified from the frozen cell pellets of cultured Huh-luc/neo-ET cells using RNeasy minikit spin columns and reagents (Qiagen, Valencia, CA) according to the manufacturer's recommended procedure. Following reverse transcription, binding sites (S1, S2, and S3, corresponding to HCV nucleotides 22 to 44 in the 5′UTR and 9387 to 9399 in the 3′UTR), were amplified by PCR (expand high fidelity; Roche, Almere, Netherlands), and S1, S2, and S3 nucleotide sequences were determined using the BigDye Terminator v1.1 cycle sequencing kit (ABI) using a cutoff of 25% for base calling of minority bases. 5′RACE (rapid amplification of cDNA ends) and sequencing analysis of the full-target HCV 5′UTR was performed by reverse transcription with a gene-specific primer located in the 5′UTR (in vitro samples) or in the HCV core region (clinical samples) (5′RACE system; Invitrogen). Sequence analysis was performed as described above. The LBCM4279 HCV isolate (genotype 1; GenBank accession number HM043170) was used as a reference sequence for 5′UTR analysis. Data from HCV nucleotides 1 to 69 (in vitro samples) or 1 to 341 (clinical samples) (the 5′UTR) are reported and comprise the S1 (nucleotides 23 to 28) and S2 (nucleotides 38 to 43) binding sites.

Open in a separate window

FIG 3

HCV replicon genotype 1b cells were serially passaged in the presence of G418 alone (ctrl) or G418 with miravirsen (MIR), SPC4729 (oligonucleotide negative control), or telaprevir (TVR) for time periods of up to 35 days (TVR) or 148 days (MIR, SPC4729, or G418 alone) in fixed or escalating concentrations at a multiple (x) of the EC₅₀ concentrations of MIR (for MIR and SPC4729) or TVR. Colony formation, drug susceptibility assays, and nucleotide changes (*) were assessed throughout the study (d, days).

Sequence analysis of HCV clinical isolates.

The selection and treatment of human subjects has been described previously (33). Briefly, a total of 36 subjects with chronic HCV genotype 1 infection were randomly assigned to receive five weekly subcutaneous injections of miravirsen at doses of 3 mg, 5 mg, or 7 mg per kilogram of body weight or placebo over a 29-day period. Subjects were monitored (HCV RNA levels and safety information) for 18 weeks after receiving the first dose of miravirsen. Amplification and sequence analysis of the entire 5′UTR (nucleotides 1 to 341) were performed by 5′ RACE as described above on representative samples from six subjects who had experienced virologic rebound (defined by an equal-to-or-greater-than 1 log₁₀ increase in HCV RNA over the nadir) and from control samples derived from seven subjects who did not experience virologic rebound following a >1 log₁₀ initial decrease in HCV RNA, had a <1 log₁₀ initial decline in HCV RNA, or were treated with placebo. Sequence analysis for S1 and S2 binding sites has been reported previously (33). The study was approved by the institutional review board or ethics committee at each participating center and was conducted in accordance with the Declaration of Helsinki, good clinical practice guidelines, and local regulations. All patients provided written informed consent before enrollment into the study.

In vitro antiviral activity of miravirsen in combination with other anti-HCV therapeutics.

Based on its unique mode of action, miravirsen was predicted not to interfere with the antiviral activity of other anti-HCV agents. To confirm this, the activity of miravirsen was evaluated when combined with other anti-HCV agents. Antiviral activity and cytotoxicity resulting from two-drug combinations of miravirsen with IFN-α2b, RBV, inhibitors of NS5B polymerase active and allosteric sites (2′-MeC and VX-222, respectively), NS3 protease (telaprevir), or NS5A (BMS-790052) were analyzed by the method of Prichard and Shipman, Jr. (36), as described in Materials and Methods. Briefly, the measured antiviral effects from each drug combination were subtracted from the antiviral effects expected if the drug interactions occurred independently (i.e., were additive). This difference then was portrayed as a three-dimensional surface-expressed dose response as a volume above (synergy) or below (antagonism) a horizontal plane which represented the level of inhibition if the drug combination was merely additive. As shown in Table 1, all two-drug combinations of miravirsen with other compounds produced total volumes calculated at the 95% confidence interval of less than 50 μM²%, indicating additive activity. The representation of these volumes as a three-dimensional dose-response surface above or below the horizontal plane representing the effects of no drug interaction is depicted in Fig. 2. In all experiments, no cytotoxicity was observed with any drug alone or in combination (data not shown).

Open in a separate window

FIG 2

Analysis of the interaction of miravirsen with IFN-α2b (A), RBV (B), VX-222 (C), BMS-790052 (D), telaprevir (E), and 2′-MeC (F) by the method of Prichard and Shipman, Jr., using Macsynergy II, v1.0 (36). Eight successive 2-fold dilutions of miravirsen were evaluated alone or in all possible combinations with five successive 2- or 5-fold serial dilutions of a second compound. Calculated independent effects were subtracted from the observed combined effects. Volumes with positive values at the 95% confidence interval indicate synergy, while volumes with negative values indicate antagonism. Data are from one representative experiment. Conc., concentration.

In vitro antiviral activity against drug-resistant HCV replicons.

In vitro experiments were conducted to evaluate the activity of miravirsen against HCV replicons constructed to contain substitutions that confer resistance to various DAA classes. Specifically, miravirsen was evaluated against HCV genotype 1b replicons containing key amino acid substitutions, rendering them resistant to the NS3 protease inhibitor telaprevir (A156T and R155K), the NS5B polymerase inhibitors 2′-MeC (S282T) and VX-222 (M423I), and the NS5A protein inhibitor BMS-790052 (Y93H). Each DAA was included as a positive control, and HCV genotype 1b replicon constructs were introduced into Huh-7 cells by transient transfection. In these assays, miravirsen demonstrated broad antiviral activity against all drug-resistant HCV variants tested with fold changes in susceptibility of less than 2-fold. In contrast, HCV replicons constructed to contain specific mutations demonstrated resistance (fold changes in susceptibility ranging from 4.4 to 45.8) distinct for each drug class tested (Table 2).

In vitro resistance studies. (i) Passage history, colony formation, and susceptibility testing.

To evaluate the ability of HCV to develop resistance to miravirsen when subjected to drug pressure, a series of in vitro serial passage studies was conducted. In this study, Huh-7 cells harboring HCV genotype 1b replicons were cultured in the presence of fixed or escalating concentrations of miravirsen for time periods of up to 148 days (Fig. 3). Concentrations ranged from 1.00 to 80 μM (approximately 2- to 160-fold the average miravirsen EC₅₀ of 0.5 μM calculated at the onset of these experiments). The NS3/4a protease inhibitor telaprevir was used as a positive control. Negative controls included SPC4729 and medium containing G418 alone. Colony formation, susceptibility assays, and sequence analyses were performed on passaged cells throughout the study.

The presence of miravirsen resulted in a decrease in the rate of cell expansion and a reduction in HCV RNA without the emergence of distinct clonal populations following the initial 25 days of passage at concentrations of up to 20-fold the average EC₅₀ (10 μM) (Fig. 4). In contrast, selection in the presence of telaprevir resulted in a decrease in the rate of cell expansion with the appearance of distinct individual resistant clonal populations (Fig. 4). Subsequently, miravirsen- and SPC4729-treated cells were cultured for an additional 10 days (35 cumulative days from the initiation of study), and antiviral activity was evaluated in each of these cell cultures. Miravirsen demonstrated comparable anti-HCV activity against cells passaged in the presence or absence of miravirsen, with EC₅₀s of 1.79 and 0.79 μM, respectively (data not shown), confirming the lack of measurable resistance and consistent with the lack of the emergence of distinct clonal populations seen at the earlier time point of 25 days of passage.

Open in a separate window

FIG 4

HCV replicon genotype 1b cells were passaged in the presence of G418 alone (medium control) or G418 with miravirsen, SPC4729 (oligonucleotide negative control), or telaprevir for 28 days in fixed concentrations at a multiple (X) of the EC₅₀ of miravirsen or telaprevir. Colony formation was assessed by staining surviving cells with crystal violet.

To continue to select for potential resistance, HCV replicon cells that had been treated with 10 μM miravirsen, 10 μM SPC4739, or G418 alone for 39 days were passaged for an additional 33 days (72 cumulative days from the initiation of study) in 2-fold increasing concentrations of miravirsen or SPC4729, up to a maximal concentration of 80 μM (160-fold the average miravirsen EC₅₀), and reevaluated for colony formation. During the conduct of these studies, cells passaged in increasing concentrations of miravirsen demonstrated a decrease in cell expansion but were never completely eliminated; the remaining cells existed as a thin patchy monolayer but not as distinct colonies (data not shown).

Since experimental conditions did not lead to the complete clearance of the HCV replicon from cells or result in the emergence of distinct resistant colonies, HCV replicon cell cultures were passaged with miravirsen at concentrations of 40 μM and 80 μM (80- and 160-fold the miravirsen EC₅₀, respectively) for an additional 27 days (99 cumulative days from the initiation of study). Distinct colonies still were not observed for cells passaged in 40 μM miravirsen, but small patches of viable cells were observed for cells passaged in 80 μM miravirsen (160-fold the EC₅₀) (data not shown). The latter cells were further cultured for 29 days in the presence of 80 μM miravirsen (128 cumulative days from the initiation of the study), and four individual HCV replicon cell clones were harvested and expanded for an additional 20 days (148 cumulative days).

Antiviral activity subsequently was evaluated for cells passaged in the presence of 40 μM miravirsen (80-fold the EC₅₀), 80 μM SPC4729, or G418 alone for 124 days and cells from clone 1 that had been passaged in the presence of 80 μM miravirsen (160-fold the EC₅₀) for 148 days in an assay using levels of HCV RNA as an endpoint (Fig. 3). In these assays, miravirsen demonstrated comparable anti-HCV activity in SPC4729- and G418-passaged cells, with EC₅₀s of 0.78 μM and 1.80 μM, respectively, but did not inhibit HCV in cells passaged in the presence of 40 μM miravirsen or clone 1 up to the highest concentration tested (320 μM; data not shown).

(ii) Sequence analysis.

To evaluate genotypic changes that might be associated with reductions in susceptibility to miravirsen, we next performed sequence analysis on stored samples from G418 control-, miravirsen-, or SPC4729-passaged cells at 39, 72, 117, and 148 days of the serial passage study (Fig. 3 and Table 3). Initially, we amplified and performed population-based sequence analysis of miR-122 binding sites (S1, nucleotides 23 to 28; S2, nucleotides 38 to 43 [28]; and S3, nucleotides 9390 to 9396 [27]) in the HCV 5′- and 3′UTRs by site-specific primed endpoint RT-PCR, followed by population-based sequencing (Materials and Methods). No nucleotide changes of significance (e.g., changes that were observed after passage with miravirsen that had not been seen following passage with SPC4729 or G418 alone) were observed in S1, S2, or S3 binding sites from cells passaged in the presence of fixed or escalating concentrations of miravirsen (data not shown). Since miR-122 has been reported to bind to nucleotides at the extreme end of the 5′UTR (29), we then used 5′RACE to amplify and obtain sequence for the distal end of the 5′UTR, including the first 20 nucleotides that had been unavailable using the standard technique (Table 3). Nucleotide changes were identified by the comparison of each sequence to the original replicon cell line. All cell lines, including those passaged in G418 alone or the control oligonucleotide SPC4729, contained a change at position 1 from a G to an A (G1A). HCV replicon cells passaged in the presence of 20 μM (40-fold the EC₅₀) miravirsen for 72 days also contained a nucleotide change from an A to a C at position 4 that was maintained following additional passages but was not detected in cells passaged with SPC4729 or G418 alone. The A4C change was not observed in earlier cells passaged for up to 39 days of treatment with 20-fold the EC₅₀ of miravirsen. This is in contrast to the levels observed following serial passage of telaprevir, where genotypic and phenotypic resistance emerged rapidly (Fig. 4 and data not shown) (40).

Clinical studies: sequence analysis.

In the previously reported phase 2 proof-of-concept clinical trial of miravirsen monotherapy given to treatment-naive patients with chronic HCV infection, no nucleotide changes in S1, S2, or S3 binding sites were detected in any subject, including those who had experienced virologic rebound (defined by a ≥1 log₁₀ increase in HCV RNA over the nadir) (33). To evaluate potential changes in the entire 5′UTR, including the first 21 nucleotides (nucleotides 1 to 341), we performed 5′RACE on representative samples from six subjects who had experienced virologic rebound (Fig. 5) and from control samples derived from seven subjects who did not experience virologic rebound following a >1 log₁₀ initial decrease in HCV RNA, had a <1 log₁₀ initial decline in HCV RNA, or were treated with placebo (data not shown). Nucleotide changes were identified by the comparison of each sequence to each subject's matched baseline control. Sequence analysis of the 5′ UTR from subjects who experienced virologic rebound indicated a nucleotide change from a C to a U at position 3 for 5 of the 6 subjects at the time of virologic rebound (Fig. 5). This mutation was not present at earlier time points during initial HCV RNA decline for two of these subjects (Fig. 5), nor was this mutation identified in control subjects (data not shown).

Open in a separate window

FIG 5

Subjects with chronic HCV genotype 1 infection were randomly assigned to receive five weekly subcutaneous injections of miravirsen at doses of 3 mg, 5 mg, or 7 mg per kilogram of body weight over a 29-day period (gray shading) (33). They were monitored for 18 weeks after randomization. Amplification and sequence analysis of the entire 5′UTR (nucleotides 1 to 341) were performed by 5′ RACE from six subjects (numbered 1 to 6) who experienced virologic rebound. Identification of the C3U nucleotide change (large open triangles) or wild-type sequences at position 3 (large open circles) are indicated. The lower limit of detection (LLOD) was 12 IU (1.08 log₁₀ IU per ml).

Phenotypic characterization of HCV replicons containing 5′UTR mutations.

To characterize the effect of the mutations found either in the in vitro selection or in the patient population on the susceptibility of HCV to miravirsen, the single mutations C3U and A4C were engineered into HCV replicons, and in vitro-transcribed RNA subsequently was electroporated into Huh-cure cells (see Materials and Methods). We were unsuccessful in generating in vitro-transcribed RNA encoding the single mutation G1A or the double mutations G1A/A4C due to the necessity of a G nucleotide at position 1 for optimal transcription by T7 RNA polymerase. Thus, only data for C3U and A4C HCV variants are presented.

We initially evaluated the replication capacity (viral fitness) of HCV variants 120 h following transfection (see Materials and Methods). The relative replication capacity of replicon sequences containing C3U and A4C substitutions in the 5′UTR were 3.0% and 34% of the level of wild-type sequences, respectively, indicating a deleterious effect of the mutations on viral fitness.

We next evaluated the susceptibility of HCV mutant replicons to miravirsen and included telaprevir and SPC4729 as controls. Overall, there was a poor dose response when testing miravirsen against the C3U HCV variant that precluded high-fidelity curve fitting and definitive EC₅₀ determinations for all but one of six experiments (Table 4 and data not shown). In that experiment, we measured an EC₅₀ of 51 μM, yielding a fold change in susceptibility of 7 compared to the EC₅₀ of wild-type virus. For some of the other experiments there was some indication of a dose response at higher concentrations of miravirsen, although in general the inhibition was incomplete, with an overall reduction in susceptibility of the virus to miravirsen. In contrast, the HCV C3U variant remained susceptible to the NS3 protease inhibitor telaprevir with a less than 1-fold change in susceptibility. We also evaluated the susceptibility of the HCV A4C variant to miravirsen; HCV replicons constructed to contain the single A4C mutation were fully susceptible to miravirsen as well as to telaprevir (Table 4).

We thank Constance Crowley for her critical review of the manuscript and Meike van der Ree for her invaluable advice and provision of clinical samples.