Potential conflict of interest: The author has received research grants from Gilead and Roche. He has served as an advisor for Abbott, Anadys, Biotica, Boehringer-Ingelheim, Bristol-Myers Squibb, DebioPharm, Gilead, GlaxoSmithKline, Idenix, Janssen-Cilag, Madaus-Rottapharm, Merck, Novartis, Pfizer, Pharmasset, Roche, Schering-Plough, Tibotec, Vertex, and Virco.
Current treatment of chronic hepatitis C virus (HCV) infection is based on the combination of pegylated interferon-α and ribavirin. The recent development of direct-acting antiviral (DAA) molecules that are active on HCV, together with in vitro and in vivo studies showing that these drugs may lead to the selection of resistant viruses if administered alone, has raised concerns that resistance may undermine therapy based on DAAs. A new standard-of-care treatment will soon be available for both treatment-naive and treatment-experienced patients infected with HCV genotype 1, based on a triple combination of pegylated interferon-α, ribavirin, and a protease inhibitor (either telaprevir or boceprevir). With this therapy, most failures to eradicate infection in treatment-adherent patients are due to an inadequate response to pegylated interferon-α and ribavirin, in the context of a low genetic barrier to resistance of first-generation protease inhibitors. This article reviews patterns of resistance to HCV DAA drugs in development, the mechanisms underlying treatment failure when these drugs are combined with pegylated interferon-α and ribavirin, the consequences of treatment failure, and possible means of optimizing future therapies that use DAAs. (HEPATOLOGY 2011;)
Onset of the AIDS pandemic in the early 1980s led to the creation of active antiviral drug discovery programs that have brought to market a large and growing number of direct-acting antiretroviral drugs with various targets and mechanisms of action. It rapidly became clear that viral resistance would be a major problem during long-term antiretroviral therapy.1 Combinations of drugs with different viral targets and no cross-resistance emerged as a valuable option for preventing human immunodeficiency virus (HIV) resistance in the long term. Multidrug-resistant viruses nonetheless emerged in some patients after several years of combination therapy.1 These viruses escaped the antiviral effects of all available drugs, leading to treatment failure, disease progression, and death. Resistance is still the principal challenge in anti-HIV therapy.
Hepatitis C virus (HCV) shares many properties with HIV: it is a highly variable virus with a quasispecies distribution, large viral populations, and very rapid turnover in the individual patient. However, unlike HIV, the HCV replicative cycle is exclusively cytoplasmic, and there is no host genome integration or episomal persistence in infected cells.2 HCV infection is therefore inherently curable. Current treatment of chronic HCV infection is based on the combination of pegylated interferon (IFN)-α and ribavirin. A sustained virological response (SVR), defined by undetectable HCV RNA 24 weeks after treatment completion, is associated with permanent cure in more than 99% of cases.3 Nevertheless, only 40%-50% of patients infected with HCV genotype 1 and up to 80% of those infected with genotypes 2 or 3 achieve an SVR with this regimen.4-6
The recent development of direct-acting antiviral (DAA) molecules active on HCV, together with in vitro and in vivo studies showing that these drugs may lead to the selection of resistant viruses if administered alone, has raised concerns that resistance may undermine DAA-based therapy. This review examines patterns of resistance to HCV DAAs in development, the mechanisms underlying treatment failure when DAAs are combined with pegylated IFN-α and ribavirin, the consequences of treatment failure, and possible means of optimizing DAA-based therapy.
In order to achieve an SVR, it is necessary (1) to shut down virus production and thereby to achieve a rapid initial decline in circulating HCV RNA; (2) to maintain viral inhibition throughout treatment; and (3) to induce a significant, slower second phase of decline in HCV RNA, leading to gradual clearance of HCV-infected liver cells, through cell death or, more often, HCV removal (Fig. 1).7 It has been hypothesized that the second-phase decline is driven by the patient's adaptive immune response in the context of sustained inhibition of virus production.7 Nevertheless, no objective findings subsequently support this hypothesis. Another hypothesis is that restoration of the intracellular innate immune response by viral inhibition plays an important role in the clearance of residual HCV genomes from these cells.
Both IFN-α and DAAs have potent antiviral properties and are able to induce a rapid first-phase decline in HCV RNA. Ribavirin addition enhances the second-phase decline induced by pegylated IFN-α, accelerating the clearance or cure of infected cells through unknown molecular mechanisms.8-10 Recent data may suggest that ribavirin could have the same effect when combined with DAAs.11 An SVR can only be achieved if the second-phase decline is gradual and treatment lasts long enough to ensure that all infected cells are cleared or cured (Fig. 1). Otherwise, HCV replication resumes shortly after treatment completion and the patient experiences a clinical relapse. Hepatitis C treatment failure is defined as the failure to achieve an SVR, i.e. a failure to eradicate HCV from the patient.
Concepts in Resistance to DAAs
Most knowledge on viral resistance to antiviral drugs comes from experience with antiretrovirals. Recent reports show that the same concepts are applicable to other viruses, including hepatitis B virus and HCV. Viral resistance to a DAA corresponds to the selection, during treatment, of viral variants that bear amino acid substitutions altering the drug target, and that are therefore less susceptible to the drug's inhibitory activity.1 These drug-resistant variants preexist as minor populations within the patient's quasispecies (the ensemble of all viral variant populations present in a given individual), as a result of the error-prone activity of the HCV RNA-dependent RNA polymerase (RdRp), the large viral populations and the short half-life of the virus in peripheral blood.12 Preexisting drug-resistant variants are rarely detected with current techniques prior to therapy, because the amino acid substitution(s) that confer resistance also generally reduce replicative capacity in the absence of the drug. More sensitive techniques, such as ultra-deep pyrosequencing, have been used to identify such resistant variants prior to treatment.13-15
Drug exposure profoundly inhibits replication of the dominant, “wild-type” drug-sensitive viral population, and the resistant variants gradually occupy the vacant replication space. Partial resistance may allow a virus to replicate sufficiently for further mutations to accumulate, leading to stepwise decreases in drug susceptibility, albeit often at a cost of reduced replicative capacity.1 “Compensatory” or “secondary” mutations may restore the fitness of the resistant virus in vivo, allowing replication to reach near-baseline levels.1 Cross-resistance (i.e., overlapping resistance) between two antiviral drugs that target the same site or function is due to amino acid substitutions that confer reduced susceptibility to both drugs.
The degree of drug resistance can be measured in vitro. In cell-free enzyme assays, resistance is measured as the fold increase in the 50% and 90% inhibitory concentrations (IC50 and IC90, respectively), i.e., the drug concentrations that inhibit the tested enzyme function by 50% and 90%. In cell culture systems permitting full infectious cycles, and in construct-based replication systems, resistance is measured in terms of the 50% and 90% effective concentrations (EC50 and EC90). There is no consensus on the minimum fold increase needed to consider that a given amino acid substitution confers resistance. In addition, the results must be interpreted within the clinical context, because some viral variants that show low-level resistance in vitro may be more damaging in vivo than variants with higher-level resistance.
In vivo, viral resistance is influenced by three major factors:
(1)The genetic barrier to resistance, defined as the number of amino acid substitutions needed for a viral variant to acquire full resistance to the drug in question. If a single substitution is sufficient to confer high-level resistance, then the drug is considered to have a low genetic barrier to resistance, whereas the need for three or more substitutions represents a high genetic barrier. There is a low likelihood that variants bearing a large number of resistance substitutions preexist in a given patient and are fit enough to replicate at high levels when an antiviral drug is administered. Drugs with a high genetic barrier to resistance are thus less likely to be associated with clinically meaningful resistance.
(2)The in vivo fitness of the viral variant population, defined as its ability to survive and grow in the replicative environment. A selected resistant variant must have the capacity to propagate in order to fill in the replication space left vacant by the susceptible “wild-type” virus during drug exposure. Thus, a highly resistant but poorly fit virus will be less clinically significant than a less resistant but fitter virus that can replicate efficiently in the presence of the drug. Compensatory mutations may restore the fitness of a resistant variant and allow it to replicate efficiently in the presence of the drug, and even to persist after drug withdrawal.
(3)Drug exposure, defined as the drug concentration achieved in vivo relative to the IC50-IC90/EC50-EC90 values of resistant variants. If drug levels achieved in vivo are far above these IC/EC values, then resistant variants will be efficiently inhibited, even if they are far less sensitive than the wild-type virus in vitro. This explains why resistance is not always an issue with drugs that have a low genetic barrier to resistance and that select fit and resistant viral variants. Adherence to therapy is a major driver of drug exposure in a treated patient.
Resistance is usually associated with a typical “escape pattern”, with rapid recovery of pretreatment levels of viral replication, when amino acid substitutions confer a high level of resistance without impairing fitness in the presence of the drug. Viral replication may resume more gradually if the resistant virus is not very fit. Antiviral efficacy in vivo may not be affected if a resistant variant naturally replicates at low levels and/or if the drug retains partial efficacy (particularly if drug exposure is high).1,16-18
HCV Resistance to DAA Drugs in Development
Two main groups of DAAs are in clinical development: (1) inhibitors of polyprotein processing, i.e., nonstructural 3/4A (NS3/4A) serine protease inhibitors, two members of which (telaprevir and boceprevir) are close to being approved for use in combination with pegylated IFN-α and ribavirin; and (2) inhibitors of HCV replication, which comprise four drug classes: nucleoside/nucleotide analogue inhibitors of HCV RdRp; non-nucleoside allosteric inhibitors of HCV RdRp; NS5A inhibitors; and cyclophilin inhibitors.19 Table 1 lists drugs in development that have been tested in HCV-infected patients.
Table 1. HCV DAAs for Which Clinical Trial Results Have Been Published or Presented
NS3/4A protease inhibitors
BMS-650032 (Bristol-Myers Squibb)
Nucleoside/nucleotide analogue inhibitors of RdRp
Non-nucleoside inhibitors of RdRp
Resistance profiles vary considerably from one drug class to the next and, for some classes, from one drug to the next, because of different chemical characteristics, targets, mechanisms of action, pharmacological properties, and selection of HCV variants with different characteristics.
Resistance to NS3/4A Protease Inhibitors.
Peptidomimetic inhibitors of the NS3/4A protease belong to two chemical groups: macrocyclic and linear α-ketoamide inhibitors. They bind tightly to the catalytic site of the enzyme and compete with its natural substrates, the polyprotein cleavage sites, thereby inhibiting polyprotein processing, i.e., the generation of mature viral proteins. First-generation NS3/4A inhibitors such as telaprevir and boceprevir were selected for their activity on the HCV genotype 1 protease. Telaprevir also potently inhibits the HCV genotype 2 protease and has modest activity on genotype 4.20,21 Another first-generation protease inhibitor, TMC435, is active on genotype 4, 5, and 6 proteases.22 None of the first-generation protease inhibitors in development has so far shown efficacy against genotype 3.20-22
NS3/4A protease inhibitors have a low genetic barrier to resistance. They have been shown to select resistant HCV variants in vitro, in the replicon cell culture system.23-27 Three-dimensional modeling showed that these variants bore substitutions located in close vicinity to the NS3 protease catalytic triad. Changes at these positions alter the affinity of the drug for the enzyme's catalytic site and thereby attenuate its inhibitory activity.28-31 Resistance substitutions generally reduce the catalytic efficiency of the protease.27,32,33 This explains why resistant variants are rarely detected as the dominant population at baseline in patients who have never been exposed to protease inhibitors.27,32,33 However, compensatory substitutions at other positions may restore fitness without affecting the level of resistance.33
Most published clinical data on NS3/4A protease inhibitor resistance have been obtained with telaprevir. Telaprevir monotherapy selects resistant viral populations within a few days or weeks, depending on the level of drug exposure. The following telaprevir resistance substitutions have been reported, by order of increasing resistance in vitro: V36A/M/C (3.5- to 7-fold increase in the IC50 compared to wild-type sensitive virus); T54A/S (6- to 12-fold); R155K/T/Q (8.5- to 11-fold); V36A/M+R155K/T (57- to 71-fold); A156V/T (74- to 410-fold); and V36A/M+A156V/T (>781-fold).34-36In vivo fitness appeared to be the principal determinant of the replication kinetics of resistant variants during treatment.37 Variants with substitutions at position 156, which are the most resistant but least fit, were selected early during therapy. They were rapidly replaced by fitter variants bearing substitutions at positions 155, 36+155, and 36+156 at the time of virological breakthrough.34,35 Differences in resistance profiles between HCV subtypes 1a and 1b have been reported. For instance, resistance-associated amino acid substitutions at position R155 require only one nucleotide change in HCV subtype 1a isolates, but two nucleotide changes in HCV subtype 1b isolates.38 Therefore, telaprevir resistance is less frequent in patients infected with HCV subtype 1b because variants bearing substitutions at position R155 are more rarely present at treatment initiation, and other resistant variants are selected when it occurs.
Dynamic changes in viral populations generally continue after telaprevir withdrawal. The wild-type viral population grows only slowly after treatment interruption and generally takes several weeks or months to regain its dominance.34 The resistant variants probably remain present, replicating at undetectable levels. However, because HCV is incapable of episomal persistence in infected cells, resistance substitutions are not truly “archived”, and it is at least theoretically possible that resistant variants are finally cleared as a result of competition with fitter, drug-sensitive viruses.
Boceprevir and telaprevir share extensive cross-resistance in vitro, a given substitution conferring slightly different levels of resistance to the two drugs.27In vivo, substitutions so far reported to confer boceprevir resistance include those found with telaprevir with comparable kinetics, plus V170A/T and V55A, which can also be selected by telaprevir.39-42In vitro data suggest cross-resistance among all first-generation NS3/4A protease inhibitors currently in development.23,43,44 Macrocyclic inhibitors such as vaniprevir, danoprevir, TMC435, or BI201335 have been shown to also select variants bearing D168A/V/T/H substitutions.15,44
A “second-generation” protease inhibitor, MK-5172, has pan-genotype activity and, in vitro, potently inhibits most viral variants that are resistant to first-generation protease inhibitors, except those with substitutions at position 156.45 Whether such compounds will select additional resistant variants when used therapeutically remains to be seen.
Resistance to Nucleoside/Nucleotide Analogues.
Nucleoside/nucleotide analogue inhibitors of HCV RdRp target the catalytic site of the enzyme, blocking the incorporation of new dideoxynucleotide triphosphates into RNA genomes in formation. They generally have pan-genotype activity, because of conservation of the RdRp catalytic site among the different HCV genotypes and subtypes.
The 2′-methyl nucleosides such as RG7128 have been reported to select substitutions at position S282, in close vicinity to the enzyme's catalytic site.46 The S282T substitution has been reported to confer a three- to 6-fold loss of in vitro sensitivity to PSI-6130, the active derivative of RG7128.46 This substitution results in a moderate loss of antiviral activity but in a large reduction in replicative capacity.47-49 Therefore, expansion of viral populations bearing substitutions that confer resistance to nucleoside analogues is expected to occur more rarely and less rapidly during treatment. Indeed, RG7128 monotherapy at various doses for a few weeks was found not to select resistance-associated substitutions.50 A weak 2′-methyl nucleoside inhibitor, valopicitabine (development of which was halted because of gastrointestinal toxicity), was shown to select variants bearing amino acid substitutions at position 282 after an average of 14-16 weeks of treatment.51
Resistance to Non-Nucleoside RNA-Dependent RNA Polymerase Inhibitors.
A number of non-nucleoside RdRp inhibitor families have been identified. These molecules essentially target four allosteric sites at the surface of the RdRp. Binding to their target site appears to induce conformational changes in RdRp, altering its catalytic capacity and reducing the synthesis of new viral genomes. Non-nucleoside inhibitors in development have a low genetic barrier to resistance. In vitro, these molecules have been shown to select a variety of resistant variants, depending on the class of drug, its target site, and its mechanism of action.39,49,52-61 Three-dimensional modeling showed that these substitutions were generally located in close vicinity to the allosteric binding site of the drug, and that changes at these positions reduced its binding affinity. Cross-resistance has been described between drugs targeting different sites. The corresponding amino acid substitutions probably induce conformational changes that impact the binding affinity of the other class of drugs.
Few in vivo resistance data have been reported with non-nucleoside RdRp inhibitors administered alone, because proof-of-concept studies of antiviral efficacy are generally limited to 3-5 days. Previously, HCV-796, a benzofuran targeting the palm 2 domain of the RdRp (finally dropped because of hepatic toxicity when combined with pegylated IFN-α and ribavirin) was administered for 2 weeks. Most patients responded initially but relapsed after a few days, because of the selection of variants bearing amino acid substitutions principally located at position C316, close to the target site.62 Filibuvir is a non-nucleoside inhibitor that targets the thumb 2 domain of the RdRp and is in development phase 2. Resistance to filibuvir is characterized by the selection of variants bearing a substitution at position M423, which was shown to interfere with filibuvir binding to its target site.63
Resistance to NS5A Inhibitors.
NS5A inhibitors select resistant HCV variants bearing amino acid substitutions in the NS5A protein. BMS-790052 has been shown to bind specifically to domain I of NS5A and to potently inhibit HCV, with EC50s of the order of 10-50 pmol in the replicon model, and pan-genotype antiviral activity.64 Its mechanism of action is unclear, however, as the NS5A protein does not have an enzymatic function and its role in the HCV replicative cycle, although crucial, remains obscure. BMS-790052 has a low genetic barrier to resistance. In vitro, it selects viruses bearing single amino acid substitutions at position M28, Q30, M21 or Y93, that confer high-level and lower-level resistance to subtypes 1a and 1b, respectively.64 So far, no selection of resistant HCV variants has been reported when BMS-790052 was used alone, but exposure was very short. In contrast, virological breakthrough has been reported when BMS-790052 was administered together with a first-generation protease inhibitor, but without IFN-α or ribavirin, emphasizing the low genetic barrier to resistance of this class of drugs and of the combination.65
Resistance to Cyclophilin Inhibitors.
Cyclophilin inhibitors are cyclosporine A analogues lacking the immunomodulatory properties of cyclosporine A. They specifically bind to cyclophilins and thereby inhibit their peptidyl-prolyl cis–trans isomerase activity. Because cyclophilins play an important (although not fully understood) role in the HCV replicative cycle by interacting with NS5A and the RdRp at the level of the replication complex, cyclophilin inhibitors potently inhibit HCV replication. Although its target is a host cell protein and not a viral component, resistance to alisporivir (DEBIO-025) appeared to be mostly associated with amino acid substitutions in the NS2 and NS5A coding regions in vitro, such as D320E in NS5A.66,67 However, the level of resistance conferred by these substitutions was low. Ongoing clinical trials will show whether mutants that bear these substitutions are actually selected in vivo and replicate at high levels.
Reasons and Mechanisms for Treatment Failure During Triple Combination Therapy With Pegylated IFN-α, Ribavirin, and a Protease Inhibitor
Resistance to antiviral drugs is classically prevented by combining several drugs with potent antiviral activity and no cross-resistance. HCV resistance to DAAs is significantly less frequent when one of these drugs is administered in combination with pegylated IFN-α or with both pegylated IFN-α and ribavirin68-70 (Fig. 2). The triple combination of pegylated IFN-α, ribavirin and a protease inhibitor (telaprevir or boceprevir) will soon become the standard-of-care therapy for treatment-naive and treatment-experienced patients with HCV genotype 1 infection.
The results of phase 2 and 3 clinical trials indicate that the triple combination of pegylated IFN-α, ribavirin and a protease inhibitor (telaprevir or boceprevir) fails to eradicate HCV infection in approximately 20%-30% of treatment-naive and 50%-60% of treatment-experienced patients.41,69-76 Different patterns of treatment failures have been reported, including nonresponse (HCV RNA never becomes undetectable on therapy), breakthrough on treatment and relapse after treatment cessation. Substantially higher failure rates are expected when these therapies are used in more difficult-to-treat populations, such as patients with unfavorable genetic markers of IFN-responsiveness and specific subgroups such as African Americans and null responders to prior therapy with pegylated IFN-α and ribavirin, or patients who have not yet been included in clinical trials, including patients with advanced liver disease, liver transplant recipients, HIV-coinfected individuals, hemodialysis patients, or immunosuppressed patients.
In treatment-adherent patients, failure of the triple combination of pegylated IFN-α, ribavirin and a protease inhibitor to eradicate HCV infection results primarily from an inadequate response to pegylated IFN-α and ribavirin, which leads to uncontrolled outgrowth of resistant variants selected by the protease inhibitor. Indeed, phase 2 and 3 clinical trials show that the outcome of triple combination therapy strongly depends on the ability of pegylated IFN-α and ribavirin to induce a sufficiently strong antiviral response in host cells, as detailed below and in Tables 2 and 3.
Table 2. Rates of Treatment Failure (Non-SVR) in Phase 2 and 3 Clinical Trials of Combination Therapy With Pegylated IFN-α2b, Ribavirin, and Boceprevir, According to the Log10 HCV RNA Decline During a 4-Week Lead-In Phase With Pegylated IFN-α2b and Ribavirin Dual-Agent Therapy
HCV RNA Decline During Lead-In (log10 IU/mL)
Treatment Failure Rate (%)
SPRINT-1: patients in the 28-week arm received 24 weeks of the triple combination after the 4-week lead-in phase; patients in the 48-week arm received 44 weeks of the triple combination after the 4-week lead-in phase (41). SPRINT-2: results for the non-black cohort; patients in the BOC/RGT (response-guided therapy) arm received 24 weeks of the triple combination after the 4-week lead-in phase if HCV RNA was undetectable (<9.3 IU/mL) at weeks 8 and 24 of therapy (weeks 4 and 20 of boceprevir administration), or 44 weeks of the triple combination after the 4-week lead-in phase if it was not; patients in the BOC/PR48 arm all received 44 weeks of the triple combination after the 4-week lead-in phase.74 RESPOND-2: patients in the BOC/RGT (response-guided therapy) arm received 32 weeks of the triple combination after the 4-week lead-in phase if HCV RNA was undetectable (<9.3 IU/mL) at week 8 of therapy (week 4 of boceprevir administration), and 44 weeks of the triple combination after the 4-week lead-in phase if it was not; patients in the BOC/PR48 arm all received 44 weeks of the triple combination after the 4-week lead-in phase.72
SPRINT-1 (phase 2, treatment-naive)
28-week arm (n = 103)
48-week arm (n = 103)
Undetectable (<15 IU/mL)
SPRINT-2 (phase 3, treatment-naive)
BOC/RGT (n = 368)
BOC/PR48 (n = 366)
RESPOND-2 (phase 3, treatment-experienced)
BOC/RGT (n = 162)
BOC/PR48 (n = 161)
Table 3. Rates of Treatment Failure (Non-SVR) in Rollover Study 107 (Phase 2) and the REALIZE Trial (Phase 3) of Pegylated IFN-α2a, Ribavirin, and Telaprevir in Treatment-Experienced Patients, According to Their Response to a Prior Course of Pegylated IFN-α and Ribavirin Dual-Agent Therapy
In rollover study 107, patients included in the control arm of the three PROVE studies were retreated with the triple combination of pegylated IFN-α2a, ribavirin, and telaprevir for 12 weeks, followed by 12 or 36 weeks of pegylated IFN-α2a and ribavirin dual-agent therapy, depending on whether they experienced an extended rapid virological response (eRVR: undetectable HCV RNA at weeks 4 and 12).77 In the REALIZE trial, the preliminary results of which have been shown only in a press release, treatment-experienced patients received the triple combination of pegylated IFN-α2a, ribavirin, and telaprevir for 12 weeks, preceded or not by a lead-in phase of 4 weeks with pegylated IFN-α2a and ribavirin dual-agent therapy and followed by pegylated IFN-α2a and ribavirin dual-agent therapy until week 48; pooled results from the two telaprevir-containing arms are shown, because results for the individual arms have not yet been reported. NA, not applicable.
Null response: <1.0 Log10 HCV RNA decline at week 4 of therapy or <2.0 Log10 HCV RNA decline at week 12.
Partial response: ≥2.0 Log10 HCV RNA decline at week 12 of therapy, but detectable HCV RNA at week 24.
Breakthrough: detectable HCV RNA after achieving undetectable HCV RNA.
Relapse: HCV RNA undetectable at the end of treatment, becoming detectable after treatment completion.
Rollover 107 (n = 117)
REALIZE (n = 530)
In the phase 2 SPRINT-1 trial of pegylated IFN-α2b, ribavirin and boceprevir, the patients were randomized to treatment arms with and without a “lead-in” phase, consisting of pegylated IFN-α2b plus ribavirin for 4 weeks, before adding boceprevir for an additional 24 or 44 weeks. Treatment failure rates are shown in Table 2 according to the reduction in HCV RNA during the 4-week lead-in phase.41 The probability of failing to achieve an SVR during triple combination therapy was highest when the HCV RNA level had been reduced by less than 1.5 Log10 IU/mL, and lowest when it had been reduced by more than 4.0 Log10 IU/mL at week 4 of the lead-in phase, regardless of the total treatment duration.41 In the phase 3 SPRINT-2 and RESPOND-2 trials of pegylated IFN-α2b, ribavirin and boceprevir in treatment-naive and treatment-experienced patients, respectively, failure rates were also significantly higher in patients who had less than a 1.0 Log10 IU/mL HCV RNA decline during the lead-in phase than in patients who had a stronger decline (Table 2).72,74
Concordant results were reported with telaprevir in rollover study 107, in which patients included in the control arms of the three phase 2 PROVE studies who had received pegylated IFN-α and ribavirin alone and had not cleared the infection were retreated with the triple combination of pegylated IFN-α2a, ribavirin and telaprevir. The treatment failure rates (Table 3) strongly depended on the response to pegylated IFN-α and ribavirin during the first course of therapy.77 Similarly, in REALIZE, a phase 3 trial assessing the efficacy and safety of pegylated IFN-α2a, ribavirin and telaprevir in nonresponders to a first course of pegylated IFN-α and ribavirin, final outcome strongly depended on the response to the first course of therapy (Table 3).
In patients in whom treatment failed to eradicate the infection and who experienced a breakthrough during treatment or a relapse after the end of therapy in the phase 2 and 3 clinical trials of telaprevir or boceprevir, protease inhibitor-resistant viral populations were dominant at the time of relapse in most if not all cases.41,69,78
A number of factors involved in the inadequate response to pegylated IFN-α and ribavirin have been described.79,80 Recently, an important role for the host's genetic background in IFN-responsiveness has been demonstrated by the observation of a strong statistical relationship between the so-called “IL28B genotype” and the virological response to pegylated IFN-α and ribavirin.81-84
Implications of Treatment Failure With a Triple Combination of Pegylated IFN-α, Ribavirin, and a Protease Inhibitor
Failure to eradicate HCV infection with a triple combination of pegylated IFN-α, ribavirin and a protease inhibitor, and the resulting outgrowth of protease inhibitor-resistant viral populations in these patients, raises two important questions: will treatment failure alter the natural history of HCV-related liver disease? Will selection of resistant viruses compromise the patient's chances of being cured of the infection by a future treatment?
Consequences of treatment failure on the outcome of liver disease.
Fears were recently raised that treatment failure with the triple combination of pegylated IFN-α, ribavirin and a protease inhibitor might be associated with accelerated progression of liver disease. Failure of triple combination therapy to eradicate HCV is characterized by a reincrease in viral replication to near-baseline levels, but the rate of disease progression appeared similar to that observed before treatment, at least in the short to medium term (approximately 3 years of follow-up thus far). This is explained by the fact that the circulating HCV RNA level does not correlate with the severity of liver disease, which progresses slowly. Because HCV is not cytopathic, there is no reason to believe that HCV variants bearing one or several substitutions in their protease would be more aggressive for the liver than the variants that predominated before therapy. Because liver disease progression is essentially due to the local immune response targeting infected hepatocytes, and as triple-drug treatment failure has not been associated with ALT flares, it is unlikely that selection of protease inhibitor-resistant variants that have been present for years as minor viral populations could lead to a shift in immunodominance resulting in a strong intrahepatic cellular immune response and associated production of proinflammatory cytokines that would accelerate liver disease progression. Thus, although long-term follow-up studies of these patients with clear clinical and histological endpoints are needed to draw firm conclusions, there is currently no reason to think that failure to clear HCV with the triple combination of pegylated IFN-α, ribavirin and a protease inhibitor is harmful. Patients in whom this regimen has failed should be reassured, as many other therapeutic options will be available in the near future. They should nevertheless continue to be followed for disease progression and retreatment options as clinically indicated.
Implications of Treatment Failure for Future DAA-Based Therapy.
Recently, preliminary results of a follow-up study of 56 patients included in the phase 2 and 3 clinical trials of pegylated IFN-α2a, ribavirin and telaprevir and in whom this treatment failed to eradicate HCV have been presented. Most of them harbored dominant telaprevir-resistant viral populations at the time of breakthrough or relapse. After a median follow-up of 25 months (range 7-36 months), 89% of them had lost the resistant virus, whereas the wild-type, telaprevir-sensitive virus had returned the dominant population.78,85 Similar results have been reported with boceprevir.86 Nevertheless, the techniques used lacked sensitivity and more precise information on the dynamics of drug-resistant viral populations should emerge soon with the use of highly sensitive technologies such as ultra-deep pyrosequencing.13-15
Whether selection of viral variants that are resistant to telaprevir/boceprevir during triple combination therapy will impact future treatment of these patients is a matter of debate. Protease inhibitor-resistant viruses are naturally relatively fit and may have acquired more fitness during protease inhibitor administration. In theory, if variants resistant to the protease inhibitor remain the dominant viral species at the time of retreatment, the use of a first-generation protease inhibitor with extensive cross-resistance with telaprevir and boceprevir in combination with other drugs is not indicated. Indeed, persistence of TMC435-resistant viral variants in previously exposed patients was reported to be associated with reduced antiviral activity of TMC435 on retreatment.87 In contrast, if the dominant viral population is wild-type, there is no theoretical contra-indication to the use of a combination of drugs that includes a first-generation protease inhibitor, provided that the other drugs have a potent antiviral effect and no cross-resistance with first-generation protease inhibitors (i.e., they are active against persisting minor protease inhibitor-resistant viral populations). However, as triple combination treatment failure is generally due to an inadequate antiviral effect of pegylated IFN-α and ribavirin, retreatment of these patients can be envisaged only if the effect of pegylated IFN-α and ribavirin has been optimized and/or several DAAs with different targets and mechanisms of action are available for use in a quadruple combination with pegylated IFN-α and ribavirin or an IFN-free regimen (see below).
Options to Prevent Treatment Failure With DAAs
Prediction of Triple Combination Treatment Failure.
Monitoring of on-treatment viral kinetics has been useful to tailor the length of treatment with pegylated IFN-α and ribavirin. Based on the design of phase 3 clinical trials, response-guided therapy will also be used to tailor the duration of triple combination treatment.72-75,84 However, a majority of patients respond quickly to the potent antiviral effect of protease inhibitors and rapidly have undetectable HCV RNA. Thus, HCV RNA kinetics is less informative than with pegylated IFN-α and ribavirin dual-agent therapy.
Given the importance of the antiviral response to pegylated IFN-α and ribavirin on the final outcome of therapy, it may be appropriate to assess this response prior to starting the DAA, particularly in prior nonresponders to pegylated IFN-α and ribavirin. The lead-in phase, during which only pegylated IFN-α and ribavirin are administered for 4 weeks, is particularly informative in this respect, as it has been shown to strongly predict the likelihood of an SVR.41,72,74 How this information should be used in practice remains a matter of debate. Patients with an inadequate response to pegylated IFN-α and ribavirin should not be excluded from therapy with telaprevir or boceprevir, as they still have a 20%-30% chance of an SVR. Nevertheless, prospective clinical trials are urgently needed to assess strategies designed to substantially improve SVR rates in this group of patients.
As the lead-in phase may be cumbersome in clinical practice, identification of baseline parameters that closely correlate with IFN-ribavirin responsiveness and have strong predictive value for the outcome of triple combination therapy could prove particularly useful for tailoring therapy in future. Further analysis of phase 3 trials with boceprevir and telaprevir is likely to provide helpful information on baseline predictors of SVR to triple combination treatment, particularly as a function of the response to IFN-α and ribavirin in trials including a lead-in phase. There is little doubt that parameters such as the IL28B genotype and the serum IFN-γ-induced protein 10 (IP-10) level will be important,88 together with classical predictors of IFN-ribavirin responsiveness.79,80 It is unlikely, however, that any of these parameters alone will have sufficient predictive value to guide treatment decisions. Therefore, predictive algorithms combining several baseline parameters need to be developed and prospectively validated for practical clinical decision-making.
Prevention of HCV Treatment Failure With DAA-Containing Regimens.
Prevention of DAA-resistant virus outgrowth is based on the use of combinations of potent antiviral drugs with no cross-resistance. Several strategies can be envisaged in the next few years to prevent HCV treatment failure and improve SVR rates. For the next 3 to 4 years, the only option for patients who do not respond well to IFN and ribavirin will be to improve the antiviral efficacy of these two drugs combined with telaprevir or boceprevir. High doses of pegylated IFN-α have been shown to be able to restore a strong response to IFN-α in an IL28B genotype-dependent manner in a substantial proportion of patients who do not respond to pegylated IFN-α and ribavirin.89 High-dose pegylated IFN-α for the full duration of therapy should now be tested in combination with ribavirin and a protease inhibitor in this population. The addition of a second DAA to pegylated IFN-α, ribavirin, and a protease inhibitor could raise the genetic barrier to resistance of the DAAs and allow more patients with a modest response to pegylated IFN-α and ribavirin to achieve an SVR. This possibility must be tested. However, the second DAA must be chosen carefully, as recent data suggest that the genetic barrier of combination therapy with a first-generation protease inhibitor and a non-nucleoside RdRp inhibitor or an NS5A inhibitor is not higher than that of protease inhibitor monotherapy.11,65,90 In contrast, no breakthroughs occurred during a 2-week course of a nucleoside analogue and a protease inhibitor, because of the high barrier to resistance of this combination.91 Nevertheless, the benefit of a quadruple combination over a triple combination including a DAA with a high genetic barrier to resistance (such as a nucleoside analogue RdRp inhibitor or a cyclophilin inhibitor) should be prospectively assessed.
The use of several DAAs without IFN-α is another option. Key questions will include the number of drugs that should be used together, how they should be combined to achieve a high genetic barrier to resistance, and their ability to trigger a sustained second slope of viral decline that will eventually lead to viral eradication, especially in patients in whom a triple combination of pegylated IFN-α, ribavirin and a protease inhibitor failed to eradicate HCV. Mathematical modeling suggests that at least three drugs should be used,92 but the final number will depend on their modes of action and the likelihood that HCV variants bearing substitutions in different regions of the genome conferring resistance to the different classes of drugs are present in the same strain. Recent data suggest that, even in the absence of IFN-α, ribavirin accelerates the second slope of viral decline, prevents relapses and may eventually increase SVR rates, as it does when combined with pegylated IFN-α, through molecular mechanisms that remain to be identified. Anecdotal cases of HCV eradication (one chimpanzee and two patients) with DAAs and without IFN-α have been reported.93-95 It remains to be seen how many, and which patients may benefit from IFN-free regimens.
A new standard-of-care treatment will soon be available for both treatment-naive and -experienced patients infected with HCV genotype 1, based on a triple combination of pegylated IFN-α, ribavirin, and either telaprevir or boceprevir. With this therapy, most failures to eradicate HCV infection are due to an inadequate response to pegylated IFN-α and ribavirin, in the context of a low genetic barrier to resistance of first-generation protease inhibitors. The same will be true with any DAA with a low genetic barrier to resistance used in combination with pegylated IFN-α and ribavirin. Such failures should not have major implications for the subsequent outcome of HCV-related liver disease but may influence the choice and results of the next line of treatment, depending on the availability of new drugs and regimens. Accurate prediction of triple combination treatment failure and effective preventive measures should help minimize pointless drug exposure and the selection of resistant viruses. As HCV infection is inherently curable, the development of new potent combinations of drugs with a high barrier to resistance will likely lead to effective control of HCV in parts of the world where such therapies are affordable. Elsewhere, an effective prophylactic vaccine will remain the only viable option.