| |
Week 48 resistance surveillance in two phase 3 clinical studies of adefovir dipivoxil for chronic hepatitis B
|
| |
| |
Hepatology July 2003, Volume 38, Number 1
Christopher E. Westland, Huiling Yang, William E. Delaney IV, Craig S. Gibbs, Michael D. Miller, Carol L. Brosgart, 437 and 438 Study Teams. Gilead Sciences, Inc., Foster City, CA
"......We performed a comprehensive evaluation of the potential development of resistance in all patients treated with ADV or placebo for 48 weeks in 2 phase 3 clinical studies. This surveillance program consisted of 3 components: (1) sequencing of the complete HBV-RT domain from all patients, (2) phenotypic analysis of conserved site substitutions that emerged in ADV-treated patients, and (3) phenotypic analysis of HBV clones isolated from ADV-treated patients with 1 log10 or greater increases in HBV-DNA levels..... ADV appears to have a unique resistance profile. In contrast to the 14% to 32% resistance rate in chronic hepatitis B patients receiving 1 year of lamivudine therapy, no adefovir resistance mutations were identified in 467 ADV-treated patients after 48 weeks of treatment with ADV in the 2 phase 3 studies. These results confirm those from smaller groups of patients during phase II studies...... Although no adefovir resistance mutations were identified in chronic hepatitis B patients after 48 weeks of adefovir dipivoxil therapy in these analyses, drug resistance is likely to emerge to any antiviral compound during extended monotherapy. To this end, a prospective 5-year study that includes genotypic surveillance and analysis of patients with virologic rebound currently is ongoing. Preliminary information recently became available from a cohort of patients treated with ADV for 2 years. A unique adefovir resistance mutation (rtN236T) was identified and is currently being characterized in vitro and in vivo..."
Abstract
Seven hundred nucleoside treatment-naive patients were enrolled in two phase 3 trials of adefovir dipivoxil (ADV) for the treatment of chronic hepatitis B. To monitor for the emergence of potential adefovir resistance mutations over the first 48 weeks, all intent-to-treat patients (467 ADV-treated and 228 placebo patients) were included in a prospectively defined, treatment-blinded, virology substudy. The study protocol mandated genotypic analysis for all patients with detectable hepatitis B virus (HBV) DNA by Roche Amplicor polymerase chain reaction (PCR) at baseline and week 48, and in vitro phenotypic analyses for patients with conserved site substitutions in HBV polymerase or 1.0 log10 or greater increase in HBV DNA from nadir. Paired sequences of the entire HBV reverse transcriptase were obtained for 271 ADV-treated and 227 placebo patients by using a sequencing method that detects down to 30% of minor species present within mixtures. Four substitutions (rtS119A, rtH133L, rtV214A, and rtH234Q) developed once each at conserved sites in HBV polymerase in 4 ADV-treated patients. Seven conserved site substitutions developed in 6 placebo patients. HBV mutants encoding the 4 substitutions that emerged in ADV-treated patients remained fully susceptible to adefovir in vitro. Furthermore, these 4 ADV-treated patients had HBV-DNA reductions of 3.3 to 5.9 log10 copies/mL by week 48 with no rebound. All other substitutions occurred at very low frequencies (<1.6%) at polymorphic sites and were not associated with HBV-DNA increases in patients or adefovir resistance in vitro. In conclusion, no adefovir resistance mutations were identified in a large group of chronic hepatitis B patients treated with ADV for 48 weeks.
Background
Hepatitis B virus (HBV) is a major cause of chronic liver disease with an estimated 400 million chronic carriers worldwide. Treatment of chronic hepatitis B aims to achieve sustained suppression of HBV replication, normalization in alanine aminotransferase levels, loss of serologic markers of hepatitis B e antigen or hepatitis B surface antigen, and improvement in liver histology. Before the recent regulatory approval of adefovir dipivoxil for the treatment of chronic hepatitis B in the United States, lamivudine was the only oral agent indicated for this disease. However, a major drawback to long-term lamivudine therapy is the emergence of lamivudine-resistant HBV mutants in a significant proportion of patients. Resistance to lamivudine occurs in 16% to 32% of patients after 1 year of monotherapy and the incidence progressively increases to 67% at 4 years. Single mutations (rtM204V or rtM204I) in the tyrosine-methionine-aspartate-aspartate (YMDD) motif of HBV polymerase were shown to be less fit than wild-type HBV during in vitro studies. However, in the majority of patients, YMDD mutations are accompanied by compensatory mutations. Secondary mutations of rtL180M and rtV173L have been observed in greater than 80% of lamivudine-resistant patients and these mutants are able to restore replication fitness of YMDD mutant HBV in vitro. Furthermore, long-term data from patients harboring lamivudine-resistant HBV indicates that viral replication can return to pretreatment levels, and is sufficient to drive disease progression in many patients. In addition, lamivudine-resistant HBV has been reported to be transmissible.
HBV is a dynamic virus that constantly undergoes changes in its genome during the course of chronic infection. The mutation rate of hepadnaviruses has been estimated at 2 ´ 10Ð4 nucleotide substitutions/site/year. Spontaneous viral mutations result in diversity of HBV sequences and define the natural polymorphisms of HBV. In HBV polymerase, some sites have been highly conserved during evolution, while others can tolerate variability (i.e., polymorphism). Amino acids at conserved sites are likely to play more crucial roles in protein function and viral fitness and, thus, are maintained due to a higher competitive advantage over substitutions at these sites. To date, all known human immunodeficiency virus (HIV) and HBV resistance mutations to nucleoside/nucleotide analogues have been located at conserved sites in their respective polymerases.
Emergence of drug-resistant virus depends on the interaction of viral factors (e.g., mutation rate, replication fitness, tropism), host factors (e.g., genetic background, immune pressure, target cell availability), and drug factors (e.g., chemical structure, antiviral potency, metabolism). With currently available oral HBV therapies, long-term dosing is required in most patients. Therefore, it is critical for drugs to be able to suppress HBV replication for an extended period of time without inducing a significant rate of resistance.
Adefovir dipivoxil is an oral prodrug of adefovir, a nucleotide analog that possesses in vitro and in vivo activity against wild-type and lamivudine-resistant HBV. The active intracellular metabolite, adefovir diphosphate, acts as a competitive inhibitor of HBV polymerase and as a chain terminator of replicating HBV DNA. No adefovir-resistant HBV mutations have been identified in patients treated up to 136 weeks in previous phase 2 clinical studies or in vitro. In 2 phase 3 clinical studies of adefovir dipivoxil for HBV, we conducted prospectively defined, treatment-blinded, genotypic and phenotypic analyses for all patients who received at least one dose of study drug for 48 weeks to characterize the resistance profile.
Patients
Studies 437 and 438 were randomized, double-blind, placebo-controlled, phase 3 clinical studies of the safety and efficacy of adefovir dipivoxil for the treatment of hepatitis B e antigen-positive (study 437) and -negative (study 438) chronic hepatitis B. A total of 700 patients enrolled in these trials. Most of the patients were nucleoside treatment-naive patients except for 3.3% and 16.3% of the patients that had 12 weeks or less of exposure to nucleosides before enrollment in studies 437 and 438, respectively. During the first 48 weeks of the trials, 695 patients received at least one dose of study medication; these patients represent the protocol-defined intent-to-treat population (n = 228, 294, and 173 for the placebo, ADV 10-mg, and ADV 30-mg groups, respectively). Informed consent was received from all patients. The study protocols were approved by the Institutional Review Boards of all participating hospitals.
Median HBV-DNA was 7.1 log10 copies/mL in Study 438 (Hepatitis B e antigen negative) and 8.4 log10 copies/mL in study 437 (Hepatitis B e antigen positive). Median age 33 in Study 437 and 46 in Study 438.
RESULTS
The majority of week 48 samples (183 of 695, 26%) were not sequenced due to low HBV-DNA levels (<400 copies/mL), as prospectively defined. For the remaining 14 patients without paired sequencing data, 12 were due to unsuccessful PCR amplification; this was primarily attributable to low HBV-DNA levels (mean <2,000 copies/mL) in the week 48 samples. The final 2 patients had unmatched HBV sequences with nucleotide sequence divergences of 7.4% or greater between the baseline and week 48; these results were confirmed by sequencing analysis of samples collected at different time points from these patients. The sequence divergences did not result in conserved site substitutions when the week 48 sequences were compared with their corresponding baseline sequences. The observed unmatched HBV sequences in these 2 patients appeared to be caused by co-infection with multiple HBV genotypes, as recently reported by other investigators, in which there was a switch in the dominant genotype over time.
Emerging substitutions in HBV RT
A total of 267 emerging amino acid substitutions in the HBV-RT domain were observed in 128 patients. The vast majority of amino acid substitutions (93%) were observed at natural polymorphic sites in the HBV polymerase/RT domain. A total of 137 of the 267 substitutions were observed in 63 placebo patients. The remaining 130 substitutions emerged in 65 ADV-treated patients. A higher percentage of placebo patients (63 of 228, 28%) than ADV-treated patients (65 of 467, 14%) had emerging substitutions. Results from the placebo-treated patients indicated that the RT domain of the HBV polymerase gene had a natural mutation rate of approximately 0.6 amino acid substitutions/patient/year (137 substitutions per 228 placebo patients). ADV therapy reduced the mutation rate to ~0.28 amino acid substitutions/patient/year (130 of 467); this effect may be due to the potent suppression of HBV replication in treated patients.
All individual substitutions occurred at very low frequency (1.6%) in patients. The natural polymorphism rtF221Y occurred at the highest frequency (1.6%, 8 of 498 genotyped patients). However, this substitution occurred more frequently in placebo patients (7 of 228) than in ADV-treated patients (1 of 467), indicating it was not specifically selected by ADV therapy.
Conserved site substitutions
Because all known resistance mutations to nucleoside/tide HIV-RT and HBV polymerase inhibitors are reported to occur at conserved sites in the HIV-RT or HBV-RT domain, we characterized all conserved site substitutions that emerged in treated patients. A total of 11 substitutions emerged at conserved sites in the HBV polymerase/RT of week 48 isolates from 10 patients. After unblinding, 4 of the 10 patients were determined to have received ADV therapy whereas 6 had received placebo. The 4 conserved site substitutions that occurred in ADV-treated patients were rtS119A/S, rtH133H/L, rtV214A, and rtH234Q.
To evaluate the susceptibility of these mutants to adefovir, in vitro phenotypic analyses were performed by transient transfection of HepG2 cells with recombinant HBV containing the conserved-sited substitutions engineered by site-directed mutagenesis. HBV carrying the rtS119A/S, rtV214A, and rtH234Q mutations replicated in vitro and were able to be analyzed in our cell-based assay. However, rtH133L mutant HBV did not yield detectable replication signals in cell culture and therefore was evaluated in an enzymatic assay. Using cell culture-based assays, YMDD mutations typically were shown to confer over 100-fold resistance to lamivudine. However, the clinically relevant cut-off value for adefovir resistance in cell culture has not been established. Because of variability in the HBV cell culture assay, any 50% inhibitory concentration (IC50) shift of less than 5-fold relative to the IC50 of wild-type HBV generally has been considered to be not significant.
In vitro phenotypic analyses showed that the IC50 of adefovir, or the inhibition constant of adefovir diphosphate, for the 4 mutant HBV strains were within 0.3- to 3.6-fold of that for wild-type HBV. These results indicated that the conserved site substitutions did not confer resistance to adefovir in vitro. Furthermore, each of the 4 ADV-treated patients with conserved site substitutions showed 3.3 log10 or more decreases in serum HBV-DNA levels at week 48 and had no evidence of HBV-DNA rebound during the first 48 weeks of ADV therapy.
The durable reductions in serum HBV DNA observed in these 4 patients are in agreement with the in vitro data, and indicate that these 4 substitutions did not cause clinical resistance to adefovir dipivoxil.
Patients with unconfirmed increase of 1 log10 or greater in serum HBV DNA from nadir at week 48
To capture any potential case of adefovir resistance not involving conserved site substitutions or mutations outside of HBV polymerase/RT domain, all patients with significant increases in serum HBV-DNA level were analyzed. A significant increase in serum HBV-DNA level was defined prospectively as a 1.0 log10 or greater (10-fold) increase in serum HBV-DNA level from the on treatment nadir at the last HBV-DNA measurement up to 48 weeks. This 1.0 log10 magnitude was selected based on a statistical analysis of natural fluctuations of serum HBV-DNA levels in 18 placebo patients in phase 2 trials that yielded a false-positive rate of 10% in placebo patients.
Among 294 patients that received ADV 10-mg therapy, 24 (8%) patients displayed an unconfirmed (i.e., occurring at one visit only) 1.0 log10 or greater increase in serum HBV-DNA level at week 48 or last visit. Using the same definition, 63 of 227 (28%) placebo patients also showed 1.0 log10 or greater increases in serum HBV-DNA levels, which reflects the fluctuating nature of HBV replication in chronic hepatitis B disease and shows that this prospective definition is a very conservative and sensitive indicator of rebound. Of the 24 ADV-treated patients with a serum HBV-DNA increase, there was documented poor drug compliance (missed doses, missed visits, or other protocol violations) in 18 cases that may have contributed to the HBV-DNA fluctuation. Furthermore, genotypic analysis revealed that 20 of the 24 patients did not have any emerging substitutions in HBV RT. Substitutions in HBV RT were observed in the remaining 4 patients.
All of the substitutions were polymorphic substitutions that also were observed in the baseline HBV isolates of other patients who responded to ADV therapy and were therefore unlikely to be associated with a resistance phenotype. Nevertheless, baseline and week 48 HBV clones were isolated from these 4 patients and tested for in vitro susceptibility to adefovir. In vitro phenotypic assays showed that the IC50 values of adefovir for the patient HBV clones changed by less than 1.5-fold post-therapy as compared with baseline, indicating that neither these polymorphic substitutions nor emerging mutations elsewhere in the HBV genome conferred resistance to adefovir.
Discussion by authors
We performed a comprehensive evaluation of the potential development of resistance in all patients treated with ADV or placebo for 48 weeks in 2 phase 3 clinical studies. This surveillance program consisted of 3 components: (1) sequencing of the complete HBV-RT domain from all patients, (2) phenotypic analysis of conserved site substitutions that emerged in ADV-treated patients, and (3) phenotypic analysis of HBV clones isolated from ADV-treated patients with 1 log10 or greater increases in HBV-DNA levels.
DNA sequencing was used for genotypic analyses because no adefovir resistance mutations were identified previously during preclinical or phase 2 clinical studies. Adefovir also has in vitro activity against HBV bearing known lamivudine- and famciclovir-resistant mutations. Unfortunately, the in vitro selection of adefovir-resistant HBV is not currently possible due to the lack of a cell culture system that allows passage of HBV. Although the woodchuck animal model was used to select lamivudine-resistant woodchuck hepatitis virus, this model did not select the clinically relevant YMDD mutations in woodchuck hepatitis virus. Furthermore, the previously known adefovir-resistant HIV mutations (K65R and K70E) did not provide guidance for searching for adefovir-resistant HBV mutations because this region of HIV RT is conserved poorly in HBV polymerase. Therefore, we entered phase 3 clinical studies without a prior knowledge of where in the HBV genome adefovir resistance mutations might emerge. Sequencing of the entire HBV-RT domain allowed us to capture all emerging substitutions in patients who received ADV or placebo. One drawback of the sequencing method was its inability to detect minor species present in less than 30% of viral population. However, our sequencing methods were validated previously by analyzing more than 100 serum samples from hepatitis B patients who failed lamivudine therapy. YMDD mutations were detected in greater than 98% of these samples, indicating our methods were sufficient to detect genotypic changes in HBV isolates from patients undergoing antiviral therapy.
In vitro drug susceptibility testing of clinical isolates played a crucial role in defining drug resistance mutations. Results from our virology study indicated that neither increases in serum HBV-DNA levels nor the emergence of substitutions in HBV polymerase necessarily indicated drug resistance. Serum HBV-DNA levels can be affected by multiple factors such as drug compliance, natural fluctuations of HBV-DNA levels, changes in host immune pressure, drug-drug interactions, and variability of diagnostic HBV-DNA assays, in addition to the emergence of resistant virus. Substitutions in HBV polymerase can emerge in drug-treated patients but also can emerge spontaneously in untreated patients. As shown in this study, amino acid substitutions emerged in 28% of placebo patients after 48 weeks. For lamivudine, resistance mutations associated with serum HBV-DNA increases in patients conferred significant reductions in drug susceptibility in vitro. The 4 conserved site substitutions observed in ADV-treated patients in this study neither conferred reduced susceptibility in vitro nor were associated with serum HBV-DNA increases. Therefore, we concluded that these substitutions were not adefovir resistance mutations, but instead may have been previously unrecognized polymorphic sites. Indeed, retrospective analysis of the 694 baseline HBV sequences from this study showed that all 4 conserved site substitutions also were present in the baseline samples (rt119A, rt133L, rt214A, and rt234Q were observed in 3, 8, 8, and 2 of the 694 baseline samples, respectively).
Site-directed mutagenesis has been used to study the effect of point mutations on drug susceptibility in laboratory HBV strains with the caveat that mutations were removed from their natural genetic context. To analyze HBV substitutions in their natural genetic background, we developed a novel in vitro phenotypic assay that permitted efficient phenotypic analysis of full-length patient-derived HBV isolates. Of the 294 patients who received ADV 10-mg daily for 48 weeks, only 4 patients had both emerging substitutions (all at polymorphic sites) and 1 log10 or greater increases in serum HBV-DNA levels. In vitro phenotypic analyses of these patient HBV clones showed no change in the adefovir susceptibility between pre- and post-therapy isolates, indicating that neither the observed polymorphic RT substitutions nor substitutions in other regions of the HBV genome were associated with resistance to adefovir. The observed HBV-DNA increase in these 4 ADV-treated patients were likely caused by other reasons such as poor drug compliance or natural fluctuation of HBV-DNA levels.
ADV appears to have a unique resistance profile. In contrast to the 14% to 32% resistance rate in chronic hepatitis B patients receiving 1 year of lamivudine therapy, no adefovir resistance mutations were identified in 467 ADV-treated patients after 48 weeks of treatment with ADV in the 2 phase 3 studies. These results confirm those from smaller groups of patients during phase II studies.
There are 2 recognized genetic mechanisms of resistance to nucleoside RT inhibitors for HIV as reviewed by Soriano and de Mendoza: (1) discriminatory mutations that favor the binding of the physiologic nucleoside substrate over that of the drug, and (2) enhanced removal of chain terminators (pyrophosphorolysis or adenosine triphosphate (ATP)-mediated removal) from replicating viral DNA. These mechanisms are also likely to apply to HBV resistance to nucleoside RT inhibitors. Molecular modeling and in vitro binding affinity testing suggested that HBV resistance to lamivudine was mainly due to steric hindrance between the side chain of valine or isoleucine at position rt204 with the sulfur atom in the oxathiolane ring of lamivudine, which is in the unnatural L-configuration (mechanism one). In another study, removal of lamivudine from lamivudine-terminated duck HBV DNA by pyrophosphorolysis was inefficient and thus removal is unlikely to be the major mechanism of lamivudine resistance in HBV. An anti-HBV agent that has a low probability of steric hindrance with mutant HBV polymerase and is difficult to remove by pyrophosphorolysis or ATP, is likely to have a low incidence of resistance.
The unique resistance profile of adefovir may be due to some inherent structural features. First, adefovir diphosphate closely resembles the natural substrate that it mimics, deoxy-ATP (dATP). A 3-dimensional model of adefovir diphosphate overlaid with dATP revealed minimal structural differences compared with those observed between lamivudine triphosphate and dCTP.
The similarity between adefovir diphosphate and dATP may limit the potential for steric discrimination by HBV polymerase mutants. Furthermore, if a mutant HBV polymerase does not bind to adefovir diphosphate, it also may not recognize its natural substrate well. Thus, such a mutant virus might be less fit. Second, unlike lamivudine, entecavir, and other L-nucleoside analogs, adefovir is an acyclic nucleotide with a flexible linker in place of the rigid sugar ring. This flexibility may allow adefovir to subvert steric hindrance created by specific HBV polymerase mutations by adopting multiple binding conformations. Third, adefovir has a phosphonate bond rather than a phosphate bond as in other nucleoside analogs. In HIV studies, tenofovir, a nucleotide with a similar structure to adefovir, was found to be less susceptible to chain terminator removal and the phosphonate bond may contribute to less efficient excision. Adefovir therefore may have the same property of resisting chain terminator removal. Finally, because adefovir already contains an -phosphorous atom, it only requires 2 phosphorylation steps to become activated, in contrast to the 3 phosphorylation steps required by nucleosides such as lamivudine. This structural feature of adefovir may lead to more efficient anabolism to the active triphosphate form and consequently improved activity in a broader range of cells such as bile duct epithelial cells, and thus reduce the HBV replication reservoirs from which resistant virus may emerge.
Although no adefovir resistance mutations were identified in chronic hepatitis B patients after 48 weeks of adefovir dipivoxil therapy in these analyses, drug resistance is likely to emerge to any antiviral compound during extended monotherapy. To this end, a prospective 5-year study that includes genotypic surveillance and analysis of patients with virologic rebound currently is ongoing. Preliminary information recently became available from a cohort of patients treated with ADV for 2 years. A unique adefovir resistance mutation (rtN236T) was identified and is currently being characterized in vitro and in vivo.
|
|
| |
| |
|
|
|
