The anti-HIV activity of entecavir: a multicentre evaluation of lamivudine-experienced and lamivudine-naive patients

Background: Entecavir, an antiviral with potent anti-hepatitis B virus activity, was recently shown to have anti-HIV activity in three patients and the ability to select for the lamivudine-resistant HIV polymerase mutation M184V in a patient with prior antiretroviral therapy.

Objectives: To further characterize entecavir's anti-HIV activity and identify risk factors for selection of the M184V.

Design: Retrospective cohort study.

Methods: We evaluated the virological characteristics of HIV and hepatitis B virus in 17 HIV-hepatitis B virus coinfected patients (10 antiretroviral therapy-naive and seven antiretroviral therapy-experienced) prior to and during entecavir monotherapy. Descriptive statistics were used to assess changes in HIV RNA and hepatitis B virus DNA. Variables associated with development of the M184V were determined by univariate analysis.

Results: Of the 17 patients, 13 (76%) demonstrated a reduction in HIV RNA by at least 0.5 log₁₀ copies/ml. Of the remaining four patients, two had the M184V detected prior to entecavir therapy and the other two had wild-type HIV. The median reduction in HIV RNA for the cohort was 1.2 log₁₀ copies/ml, which was similar in antiretroviral therapy-naive and antiretroviral therapy-experienced patients. The M184V mutation emerged in six patients receiving entecavir, including three antiretroviral therapy-naive patients. No other HIV mutations were consistently detected. Risk factors for the emergence of the M184V mutation were a decline in hepatitis B virus DNA (P = 0.04) and duration of entecavir use (P = 0.05).

Conclusion: Entecavir monotherapy in HIV-hepatitis B virus coinfected patients, including antiretroviral therapy-naive patients, has significant anti-HIV activity and can result in the development of the M184V variant. Entecavir should not be used in such co-infected patients without concomitant antiretroviral therapy.

Introduction

Approximately 400 million people globally are chronically infected with the hepatitis B virus (HBV) whereas nearly 40 million people are infected with HIV ^[1,2]. HIV-HBV coinfection is frequently observed due to shared routes of transmission, with reported figures indicating that approximately 10% of HIV-infected patients are chronically infected with HBV ^[3-7].

Current HIV guidelines recommend delaying antiretroviral therapy (ART) until CD4 cell count falls to 200-350 cells/μl due to long-term side effects and the risk of developing drug-resistant HIV ^[8]. Some HIV-HBV coinfected patients, therefore, require treatment for chronic HBV but not for HIV. The HIV-HBV group is difficult to treat because many of the anti-HBV agents have activity against HIV and, therefore, cannot be used without a full anti-HIV regimen.

Entecavir is a guanosine analogue approved for the treatment of chronic HBV ^[9-11]. On the basis of in-vitro studies indicating its lack of anti-HIV activity ^[12], entecavir was previously recommended in multiple guidelines as an agent for treatment of HBV when HIV therapy was not indicated ^[1,8,13,14] (website: http://www.aidsinfo.nih.gov/ContentFiles/AdultandAdolescentGL.pdf). We recently demonstrated that entecavir has in-vitro and in-vivo anti-HIV activity ^[15]. In that report, one of the two patients with samples available for testing selected for the lamivudine-resistant HIV variant M184V when on entecavir monotherapy. This patient had received lamivudine in the past, so it was not known if this mutation was selected from a previously archived virus or was selected de novo. In order to characterize the anti-HIV activity of entecavir better and to determine the risk factors associated with the development of M184V, we established a multicentre and international cohort study of 17 HIV-HBV coinfected patients who received entecavir in the absence of other antiretroviral agents.

Subjects and methods

Study participants

The study was a retrospective case review of 17 HIV-HBV coinfected patients from multiple sites in Australia and the United States who received entecavir as the only active HIV or HBV agent. Inclusion criteria were: HIV antibody positive at least 1 year prior to entecavir, positive for hepatitis B surface antigen (HBsAg) on at least two occasions 6 months apart, and entecavir treatment without concomitant anti-HIV or anti-HBV therapy for at least 14 days. The approval for this study was obtained from the Bayside Health Human Research and Ethics Committee, Melbourne, Australia. Oral consent was obtained from the study participants.

Data abstraction

HIV RNA, HBV DNA, and HIV-resistance test results were abstracted from medical records prior to receiving entecavir (data collected between July 2004 and July 2007). The time points selected for analysis were: last available sample prior to the introduction of entecavir, last available sample when receiving entecavir, and interim samples when receiving entecavir. In addition, all available results of HIV-resistance tests prior to receiving entecavir were collected. The time on entecavir monotherapy for the first sample available for quantification of HIV RNA ranged from 17 to 166 days (median = 54 days) and from 38 to 470 days (median = 144 days) for the last sample available for quantification of HIV RNA.

HIV RNA quantification, HIV-resistance testing, and hepatitis B virus DNA quantification

HIV RNA, HBV DNA, and HIV resistance were all determined at the participating sites' clinical laboratories using standard assays except those noted in the text. In particular, HIV RNA was quantified by the following assays: Roche Cobas HIV-1 Monitor version 1.5 [lower limit of detection (LLOD), 400 copies/ml; Roche Diagnostics, Branchburg, New Jersey, USA]; Roche Cobas HIV-1 Monitor version 1.5-Ultrasensitive (LLOD = 50 copies/ml; Roche Diagnostics); Roche Cobas AmpliPrep/TaqMan (LLOD = 40 copies/ml; Roche Diagnostics); and Bayer Versant HIV-1 RNA 3.0 (LLOD = 50 copies/ml; Bayer HealthCare, Tarrytown, New York, USA). When the HIV RNA result was more than 100 000 copies, the samples were diluted and reassayed. HBV DNA was quantified by Roche Cobas TaqMan HBV (LLOD = 170 copies/ml; 30 IU/ml; Roche Diagnostics); Roche Cobas Amplicor HBV Monitor (LLOD = 200 copies/ml; 35 IU/ml; Roche Diagnostics); or Bayer Versant HBV DNA 3.0 (LLOD = 2000 copies/ml; 357 IU/ml; Bayer HealthCare). All HBV DNA results were converted from copies/ml to IU/ml using a conversion factor of 5.6 or 5.8., depending on the particular assay used. Samples with an initial HBV DNA result over the upper limit of detection were diluted 1: 1000 and retested.

Sequencing of HIV reverse transcriptase was performed either with Viroseq HIV-1 Genotyping system version 2.0 (Celera Diagnostics, Alameda, California, USA), Phenosense GT (Monogram Biosciences, San Francisco, California, USA), vircoTYPE HIV-1 (Virco, Bridgewater, New Jersey, USA), or Trugene (Bayer HealthCare). Clonal analysis was used to assess for HIV-resistance mutations in three treatment-experienced patients (patient 13 15, and 17), and was performed as previously described ^[15].

Statistical methods

Statistical analyses were performed using Statistical package for the social sciences (SPSS) for Windows Release 14.0.02 (SPSS Inc., Chicago, Illinois, USA). The pre-entecavir values were the last available sample tested prior to the initiation of entecavir. For HBV DNA, the median was 36 days prior to entecavir initiation (range 0-118 days). For HIV RNA, it was 34 days (range 0-131 days). The decline in HIV RNA and HBV DNA values was calculated by subtracting the value prior to entecavir initiation and the lowest value on therapy (referred to as the nadir). Therefore, in some patients, there was a difference between time to nadir HBV DNA or HIV RNA and time of total entecavir exposure. For the patients with a reduction of less than 0.3 log in HIV RNA, the nadir was defined as the last available HIV RNA value on entecavir. When the HIV RNA or HBV DNA value was below the LLOD, a midway value between zero and the LLOD was used for that assay. Continuous variables were compared using the Mann-Whitney test. Categorical variables were compared using Pearson's χ² or Fischer's exact test. The relation between the decline in HBV DNA and total duration on entecavir was investigated using Spearman's rank order of correlation. All tests were two-tailed and results were considered significant at P value of less than 0.05

Results

Demographic & clinical features

A total of 17 patients fulfilled the study criteria including 10 who were naive to ART (ART-naive) and seven who had previously received ART that included lamivudine (ART experienced). One ART-naive and two ART-experienced patients were previously described (patients 10, 15, and 17) ^[15]. The median age of the cohort was 36 years, with the majority being men of Caucasian or African-American ethnicity (Table 1). The most common risk factor for HIV and HBV acquisition was men having sex with men. The median duration of known HIV infection was 8 years, which was shorter in the ART-naive compared with the ART-experienced subjects (3.5 and 15 years, respectively). The median nadir CD4+ T-cell count was 276 cells/μl (range 30-447). The ART-naive patients had a higher median CD4+ T-cell count of 356 cells/μl (range 50-450) compared with the ART-experienced patients, who had a median CD4+ T-cell count of 263 cells/μl (range 9-288). The frequency of prior AIDS-defining illnesses was similar between ART-naive and ART-experienced patients. The vast majority of patients were hepatitis B e antigen (HBeAg) positive, with 25% coinfected with hepatitis D virus (HDV) and 12% coinfected with hepatitis C virus (HCV). The only baseline characteristic that was significantly different between the groups was the increased duration of time on entecavir (defined as the time from initiation of entecavir to the discontinuation of entecavir or the end of the study period) with a median of 24.6 weeks in the ART-naive group and 47.2 weeks in the ART-experienced group (P = 0.03).

Table 1

Virological response to entecavir

Hepatitis B virus

The median pre-entecavir HBV DNA for all the subjects was 6.4 × 10⁷ IU/ml (range 4.3 × 10⁵-4.3 × 10⁹ IU/ml), and the median on-entecavir HBV DNA nadir was 4.5 × 10³ IU/ml (range 9.0 × 10⁰-5.4 × 10⁶ IU/ml) (Table 2). This represented a median reduction of HBV DNA of 3.6 log₁₀ IU/ml after a median of 152 days of entecavir (range 31-391 days). The mean HBV DNA reduction for the cohort was 4.1 log₁₀ IU/ml. The changes in HBV DNA are summarized in Table 2 and Fig. 1.

Table 2

Fig. 1

The changes in HBV DNA were not significantly different between the ART-naive and ART-experienced patients. In the ART-naive patients, the median and mean declines in HBV DNA were 3.4 log₁₀ IU/ml (range 1.6-7.3) and 4.2 log₁₀ IU/ml, respectively, after a median of 159 days of entecavir (range 31-391). In the ART-experienced patients, the median and mean declines in HBV DNA were 4.5 log₁₀ IU/ml (range 0.5-6.1) and 4.0 log₁₀ IU/ml, respectively, after a median of 144 days of entecavir (range 51-303 days).

HIV

Thirteen of the 17 (76%) patients had a reduction in HIV RNA of at least 0.5 log₁₀ copies/ml, including eight of ten (80%) ART-naive and five of seven (71%) ART-experienced patients. Of the four patients with a decline of less than 0.5 log copies/ml in HIV RNA, the two ART-experienced patients had the M184V mutation prior to entecavir initiation (patients 11 and 12), which is known to confer resistance to entecavir. The other two patients had wild-type virus (patients 4 and 6).

The median pre-entecavir HIV RNA in all subjects was 2.6 × 10⁴ copies/ml (range 7.8 × 10³-2.9 × 10⁶), whereas the median on-entecavir HIV RNA nadir was lower at 2.3 × 10³ copies/ml (range < 50-3.0 × 10⁵). This represented a median and mean overall decline in HIV RNA of 1.2 and 1.0 log₁₀ copies/ml, respectively. The median time to nadir HIV RNA was 96 days of entecavir (range 28-291). For the patients with a reduction of less than 0.3 log copies/ml in HIV RNA, time to nadir was defined as the time to the last available HIV RNA value on entecavir. Changes in HIV RNA are summarized in Table 3 and Fig. 1.

Table 3

The declines in HIV RNA were similar regardless of prior ART. In the 10 ART-naive patients, the median and mean reduction in HIV RNA was 1.3 log₁₀ (range from+0.5 to -2.3) and 1.1 log₁₀ copies/ml, respectively, after a median of 113 days of entecavir therapy (range 28-291). The seven ART-experienced patients had a median reduction in HIV RNA of 1.1 log₁₀ copies/ml (range from +0.5 to -2.3 log₁₀) after a median of 96 days of entecavir therapy (range 51-215), and the mean reduction was 0.9 log₁₀ copies/ml.

Four of these 13 patients (31%) with at least a 0.5 log copies/ml decline in HIV RNA also had a rebound of HIV RNA of more than 0.5 log₁₀ copies/ml after achieving their nadir. All four were in the ART-experienced group, of whom three had the M184V mutation detected at the time of HIV RNA rebound (patients 14, 15, and 16).

Of the seven ART-experienced patients, two (29%) had the M184V mutation detected on HIV genotypic analysis prior to entecavir initiation (patients 11 and 12). Of the remaining five patients, three (60%) developed M184V on entecavir therapy (patients 13, 14, and 15) after a median of 98 days and subsequently had viral rebound. Thus, only two ART-experienced patients (patients 16 and 17) did not select the M184V mutation. Patient 16 had received lamivudine therapy twice; briefly in 2001 as part of an ART regimen with didanosine and nevirapine, which was discontinued due to a hepatitis flare; then again in 2006, in combination with tenofovir and lopinavir/ritonavir, which was also discontinued due to hepatic flare after approximately 9 weeks of therapy. Patient 16 did not have plasma available pre-entecavir initiation for HIV reverse transcriptase sequencing, and after receiving 11 months of entecavir monotherapy, the HIV genotype did not show selection of the M184V mutation. Patient 17 had received lamivudine therapy in combination with zidovudine from 1996 to 2000. His ART was discontinued in 2000 due to noncompliance. HIV reverse transcriptase sequencing on this patient's pre-entecavir serum sample only showed the K103N mutation. The M184V mutation was not selected after 7 months of entecavir therapy.

Seven of the 10 ART-naive patients had a pre-entecavir HIV reverse transcriptase genotype, and none of them had the M184V mutation. However, on entecavir monotherapy, 2/7 (29%) of them (patients 5 and 9) selected for the M184V mutation after a median of 132 days on therapy. Of the three other ART-naive patients without a pre-entecavir HIV RT genotype, one (patient 8) had the M184V mutation detected after 260 days of entecavir therapy.

To determine the factors associated with the development of the M184V mutation, we used univariate analysis to compare those who did (n = 6) and did not (n = 7) select the variant on entecavir (Table 3). A total of 13 patients had both pre-entecavir and on-entecavir HIV reverse transcriptase resistance testing results available. A longer duration of entecavir was associated with an increased risk for developing the M184V mutation (46.0 versus 25.2 weeks, respectively, P = 0.05). In addition, HBV DNA reduction (to nadir) was significantly associated with the development of M184V (P = 0.04). Those who selected M184V had a median reduction in HBV DNA of 6.1 log₁₀ IU/ml compared with 3.0 log₁₀ IU/ml in those who did not select for M184V. To determine whether these two factors were correlated, we used Spearman's rank order of correlation. This analysis demonstrated a positive correlation between log₁₀ decrease in HBV DNA and on-entecavir time (weeks) in the entire cohort (r = 0.537, P = 0.03).

Discussion

The present study confirms the anti-HIV antiviral activity of entecavir when used as monotherapy in an HIV-infected population ^[15]. Patients receiving entecavir monotherapy had a median 1.0 log reduction in HIV RNA. These findings are in agreement with recent studies demonstrating that entecavir triphosphate can be incorporated by purified HIV reverse transcriptase in cell free in-vitro reactions. Consistent with this anti-HIV activity, the M184V variants, which have reduced the susceptibility to entecavir, emerged during entecavir monotherapy in ART-experienced and ART-naive patients ^[16].

The anti-HIV efficacy of entecavir was comparable to other nucleoside analogues such as zidovudine and lamivudine when used as monotherapy ^[17,18]. In the present study, the anti-HIV efficacy of entecavir was similar in magnitude in both the ART-naive and the ART-experienced group, suggesting that prior HIV therapy does not impair the anti-HIV activity of entecavir in this population. Interestingly, one patient had a decline in HIV RNA despite a history of the M184V mutation prior to entecavir initiation. This may have been because the predominant genotype at the time entecavir therapy was started was a wild-type form that was susceptible to entecavir. Alternatively, the decline in viremia may reflect a selection for viruses containing the M184V mutation, which has a deleterious effect on viral fitness ^[19,20]. A recent study ^[16] also showed that the M184V virus could still incorporate the substrate but at a slower rate, with lower binding affinity and lower incorporation efficiency of entecavir triphosphate compared with wild-type virus, thus, potentially also explaining some activity of entecavir against the M184V strain.

Entecavir use led to the selection of the M184V mutation in almost half of the patients (5 of 12) who did not have this mutation pre-entecavir initiation and in one additional lamivudine-naive patient who did not have a genotype prior to its initiation. In contrast to the original report, which demonstrated the selection of M184V in a lamivudine-experienced patient, we demonstrated the ability of entecavir to select for resistance in one-third of the ART-naive patients, which is consistent with a recent case report of M184V selection by entecavir in an ART-naive patient ^[21]. Although it is possible that the ART-naive patients may have been initially infected with the M184V variant and, consequently, harboured archived mutations, it is highly unlikely that this would occur at a high frequency ^[22]. Furthermore, one lamivudine-naive patient (patient 5) demonstrated the M184V mutation despite acquiring HIV well before the availability of lamivudine. In comparison, a larger proportion (60%) of the ART-experienced patients who did not have M184V preentecavir developed this mutation on entecavir therapy. This could be due to the ease of selecting a pre-existing archived mutant compared with de-novo selection. An alternative explanation is that those patients who were ART-experienced had a longer median duration of entecavir and, therefore, had more time to select for the mutation. It was also notable that in patients who experienced HIV virological rebound of 0.5 log₁₀ or more while still receiving entecavir, the majority had the M184V mutation at baseline. All of these patients had profound simultaneous reduction in HBV DNA, suggesting adherence to therapy. The fact that not all patients selected for M184V may be due to the limited replication advantage of the M184V variants, which is a consequence of the unusual dose-response curve of entecavir ^[23].

We found that selection of HIV resistance by entecavir was associated with a greater reduction in HBV DNA and longer duration on entecavir therapy. These variables were shown to be codependent and are consistent with the longer duration of entecavir use, resulting in a more profound HBV DNA drop. It is also possible that the reduction in HBV DNA may have had an indirect effect on HIV RNA by nonspecific mechanisms such as enhanced mitogen-specific and antigen-specific T-cell immunity following lamivudine treatment of HBV monoinfection, as previously described ^[24].

In several patients without the M184V mutation, entecavir administration did not result in a significant HIV RNA decline despite HBV DNA reductions. It is notable that these patients had the lowest HBV DNA reduction and a similar duration on entecavir therapy to those who did achieve a reduction in HIV RNA. It is possible that other HIV mutations prevented entecavir from having an anti-HIV effect. An alternative explanation is the potential inter-individual differences in entecavir pharmacokinetics or intracellular phosphorylation.

The present study has several limitations. First, it was a small, retrospective study; however, we believe that this is likely to be the largest cohort used to examine the anti-HIV effect of entecavir since it is no longer recommended as monotherapy in HIV-infected patients. Second, given that it was a multicentre study using different assays to measure both HIV RNA and HBV DNA, the log drop in both may be an underestimate of the true log drop, given that the most sensitive assay was not used in all patients. In addition, the duration of entecavir treatment and timing of samples was variable, with some patients having only HIV RNA and HBV DNA determined at one time point on entecavir. We may, therefore, have overestimated the time to HIV nadir and underestimated the true HIV nadir and incidence of HIV virological rebound with selection of resistance on entecavir. Finally, we did not consistently detect any other known drug-resistant HIV mutations besides M184V. In one patient (patient 14), the K70R was detected with the M184V after entecavir therapy. This was a thymidine analogue mutation and it would be highly unusual for a guanosine analogue to select this out. It is, thus, unlikely that entecavir resulted in its emergence. However, as the follow-up was limited, we cannot determine if other mutations would develop with a longer duration of therapy.

Despite the limitations of the present study, these findings clearly demonstrate that entecavir has significant anti-HIV activity and can select for the M184V mutation de novo. Therefore, its use as a single agent for HBV treatment in HIV-HBV coinfected patients should be avoided. It should, however, still be acceptable to use entecavir in a situation where HIV is well controlled by other ART and regular HIV monitoring is performed. Further, these data also support the routine assessment for undiagnosed HIV infection in all HBV-infected patients, especially before entecavir is prescribed given that it is now commonly recommended as first-line therapy for chronic HBV in this setting ^[25-27].

Acknowledgements

We thank the participants in the study; Lisa Morris (Burnet Institute), Anna Ayres and Scott Bowden (Victorian Infectious Diseases Reference Laboratory) for sample retrieval and testing; Suzanne Crowe (Burnet Institute) for supervision of specimen retrieval and HIV RNA quantification for all participants from Melbourne; and Sherilyn Brinkley, Brad Hale, Julie McCarthy and Bob Hegarty for reviewing medical histories. We also thank Nick Andrianopoulos (Monash University) for statistical advice and Greg Dore for critical reading of the manuscript.

Sponsorship: This study was partly supported by the National Institutes of Health (R01AI060449 and R01AI071820), the Alfred Foundation, and the National Health and Medical Research Council (NHMRC).

Contributors: All authors read and edited the manuscript. J.S., A.M., R.B., J.C., H.H., G.M., S.A.O., S.R.L. and C.L.T. reviewed medical histories and identified potential participants. R.F.S. and M.A.M. performed clonal analysis. J.A. was responsible for data management and statistical analysis. J.S., J.A., S.R.L. and C.L.T. analysed the data and wrote the manuscript. J.S., S.R.L. and C.L.T. designed the study.