Potential contributors to the high rate of virologic failure observed for tenofovir, abacavir, and lamivudine include a low genetic barrier to resistance for this regimen and antagonistic drug-drug interactions. To examine the second possibility, we tested combinations of abacavir, tenofovir, and lamivudine against wild-type and drug-resistant HIV-1 in vitro using peripheral blood mononuclear cells and MT-4 cells. Antagonistic interactions were not detected for any combination. If the systems examined accurately reflect the in vivo situation, antagonism does not substantially contribute to the poor efficacy of this triple combination.
From GlaxoSmithKline, Research Triangle Park, NC.
Received for publication February 1, 2005; accepted April 29, 2005.
Reprints: E. Randall Lanier, 5 Moore Drive, Research Triangle Park, NC 27705 (e-mail: email@example.com).
Two independent studies of the triple combination of abacavir, tenofovir, and lamivudine administered once a day in previously antiviral-naive HIV-infected adults have reported high rates of virologic failure (>40%) within 3 months after initiation of antiviral therapy, with selection for M184V in >90% of virologic failures and selection for K65R in >50% of virologic failures.1-3 Numerous hypotheses for the high rate of virologic failure have been postulated, all of which invoke inadequate drug levels via 1 or more mechanisms and/or the low genetic barrier to resistance for this regimen.1-4
In contrast to effective triple therapies, the genetic barrier for resistance to all the components of the abacavir/lamivudine/tenofovir combination is only 2 point mutations. Nonetheless, it is difficult to explain the poor efficacy of this regimen based solely on resistance, because early virologic failure is frequently associated with mutants that are more susceptible to tenofovir than to wild type (M184V alone). For example, the in ESS30009 trial, 44% of the nonresponding subjects had genotypic evidence at week 12 of only M184I or M184V mutations while experiencing plasma virologic rebound to near-baseline levels.3 Previous studies have shown that tenofovir is significantly more active against HIV with M184V than it is against wild type and that abacavir retains significant activity against these mutants.5,6 Conversely, the K65R mutation confers some level of resistance to all these agents. Therefore, the selective advantage of M184V over K65R evident at early virologic failure in vivo is at least superficially inconsistent with the resistance it confers. The relative reductions in viral fitness associated with M184V and K65R seem identical as measured by replication capacity in the ViroLogic (Foster City, CA) assay.7 One hypothesis that could reconcile these observations asserts that antagonism between these drugs predisposes subjects taking them to virologic rebound and selection for resistant variants.
Given that combinations of tenofovir/lamivudine/efavirenz and abacavir/lamivudine/efavirenz are highly effective, the most likely putative antagonistic interaction is between abacavir and tenofovir. We examined this hypothesis by testing double and triple combinations in vitro against HIV-1. We obtained combination activity profiles for wild-type virus and site-directed K65R and/or M184V mutants because they may be rapidly selected by abacavir/lamivudine/tenofovir and significantly alter the concentrations required for 50% inhibition (IC50s) for these drugs. Thus, it seemed possible that antagonism might be detectable in the presence of relevant mutations that was not detectable with wild-type virus.
MATERIALS AND METHODS
Compounds and Virus Strains
Abacavir and lamivudine were synthesized at GlaxoSmithKline (Research Triangle Park, NC). Tenofovir was provided by Gilead Sciences (Foster City, CA). The viruses used in this study were based on an infectious chimeric clone, pHN. The infectious clone pHN was generated by joining the second EcoRI site of HIV-1 strain HxB2 with the single EcoRI site of HIV-1 strain NL4-3. Site-directed RT mutants (containing no changes [wild type] or K65R, M184V, or K65R/M184V changes) were generated using the QuikChange Mutagenesis kit (Stratagene, La Jolla, CA) and were cloned into the infectious clone pHNRB. Infectious plasmids were calcium porated into U937 cells. Virus stocks were generated, harvested, and titered for infectivity.
Cells and Culture Assays
All culture media were obtained from Gibco BRL (Rockville, MD) unless specified. The human T-cell lymphotropic virus-transformed cell line MT-4 was cultured in RPMI 1640 medium with 10% vol/vol fetal calf serum (FCS; Hyclone, Logan, UT). MT-4 cells were infected with excess HIV-1pHN wild type, HIV-1pHN_K65R, HIV-1pHN_M184V, or HIV-1pHN_K65R/M184V (at a multiplicity of infection of 100 × tissue culture infectious dose50). Compounds were prepared in 100% dimethyl sulfoxide (DMSO) at a stock concentration of 0.01 to 0.1 M and tested at final concentrations of 0.01 to 200 μM. The experimental design and data analysis used to evaluate the effects of a combination of antivirals were as described previously.8 Drug dilutions were arranged so that every concentration of each component of the combination was tested in the presence and absence of every concentration of the other. Combination anti-HIV activity and compound-induced cytotoxicity were measured in parallel by means of an MTS-based colorimetric procedure.9
In a separate model system, combination anti-HIV activity was determined in peripheral blood mononuclear cells (PBMCs) using reverse transcriptase (RT) activity as a measure of antiviral effects. The PBMCs were prepared from HIV-1-negative blood donors (American Red Cross) by layering over Lymphocyte Separation Medium (ICN Cappel, Aurora, OH). PBMCs were cultured at 37°C in RPMI 1640 medium with 20% vol/vol FCS, 10% vol/vol interleukin (IL)-2 (Zeptometrix, catalog no. 0801017), 10 μg/mL of gentamicin, and 1.25 mg of phytohemagglutinin (PHA; Amersham Life Science, Piscataway, NJ). After 24 to 48 hours, cells were centrifuged and resuspended at a density of 2 × 106 cells/mL in RPMI 1640 medium with 10% vol/vol FCS, 10% vol/vol IL-2, and 10 μg/mL of gentamicin (growth media). The PBMCs (1 × 107) were infected with 1 mL of HIV-1. The virus concentration was adjusted based on RT measurements in counts per minute to a concentration of 0.3 to 1.0 × 106 cpm/mL. After 3 hours of virus adsorption, 40 μL of infected cell suspension was added to 210 μL of diluted test compound in 96-well plates. Plates were incubated for 11 days at 37°C, and RT levels were measured in the cell-free supernatants of each well.10
The data from triplicate assays of each combination were analyzed as previously described.8 Briefly, synergy and antagonism were defined as deviations from dose-wise additivity, which results when 2 drugs interact as if they were the same drug.11 The analysis calculates the average deviation of the data from additivity (D) and the probability that this deviation is attributable to chance. If the probability is less than or equal to 0.05, the deviation is considered significant.
The inhibition of HIV-1 by combinations of abacavir and tenofovir in PBMCs with wild-type HIV-1 plus or minus a 120-nM lamivudine overlay and the double mutant (K65R + M184V) with the lamivudine overlay is presented as isobolograms with the standard errors of both fractional inhibitory concentrations (FICs) of each combination data point plotted in Figure 1 (A, B). Combinations of abacavir and tenofovir in MT-4 cells with wild-type HIV-1 and the single and double mutant constructs (K65R, M184V, and K65R + M184V) plus a 120-nM lamivudine overlay are also shown in Figure 1 (C, D). Data using MT-4 cells were obtained for all mutants and wild-type virus in the absence of lamivudine, and data calculations are shown in Table 1. For the combination of abacavir and tenofovir in MT-4 cells with wild-type virus with or without lamivudine, the deviation from D was slightly negative and significantly different from 0. This indicated that the interaction of abacavir and tenofovir was synergistic.
Similar data were obtained in MT-4 cells in which abacavir and lamivudine were used in an equimolar mixture in combination with tenofovir, where the value for D was −0.120 and significantly different than 0, also indicating a synergistic interaction between the agents (data not shown). The data with mutant HIV-1 (K65R, M184V, and K65R + M184V) and MT-4 cells indicated additive interactions between tenofovir and abacavir in all cases (see Table 1). As expected, the data with PBMCs showed greater inherent error than the data with MT-4 cells, but the results were similar to those obtained with MT-4 cells (see Fig. 1; see Table 1). Antagonism would be indicated by a significant positive value of D; this was not observed for any combination examined in these model systems.
There was no evidence for antagonistic anti-HIV activity between tenofovir, lamivudine, and/or abacavir in vitro using MT-4 cells infected with wild-type virus or mutant constructs with K65R, M184V, or both mutations. Additionally, antagonism was not detected in PBMCs infected with wild-type HIV-1 or the dual mutant K65R/M184V. Reports from 2 other independent groups examining wild-type virus in PBMCs found no evidence of antagonism.12,13 Furthermore, recent data suggest that coadministration of abacavir and tenofovir does not lead to reduced levels of the plasma nucleoside/nucleotide or to reduced intracellular levels of the active carbovir triphosphate or tenofovir diphosphate, at least as measured in total PBMCs.14
Together, these data suggest that the explanation for poor efficacy with tenofovir/abacavir/lamivudine does not include antagonistic drug-drug interactions, although antagonism within specific infected cell subsets (eg, monocytes, macrophages, dendritic cells, specific T-cell subsets) cannot be ruled out. Notably, the strains of HIV employed here and in the other reports to date all use CXCR4. Additional studies with CCR5 tropic virus and more refined PBMC subsets could reveal antagonistic interactions not detected in this study. Another possibility to reconcile the resistance data is that cell populations exist in vivo, where lamivudine (and/or abacavir) is active (as lamivudine triphosphate and/or carbovir triphosphate), but have subinhibitory levels of active tenofovir (diphosphate). In this scenario, M184I/V might be expected to appear first as the primary mutation(s) selected by lamivudine and/or abacavir. The hypothesis of “effective monotherapy” or dual therapy despite the administration of 3 drugs could result from various mechanisms, including slower formation of tenofovir diphosphate relative to lamivudine triphosphate (and/or carbovir triphosphate), ineffective cell and/or compartment penetration, and/or cell-type-specific antagonism perhaps subsequent to inhibition of purine nucleotide phosphorylase or induction of xenobiotic membrane pumps.
These or other potentially antagonistic interactions, if extant, were not detected in the combination studies reported here or in 2 additional independent studies with PBMCs and wild-type virus.12,13 This suggests that antagonism is not a major contributor to the frequent failure of abacavir/lamivudine/tenofovir to control HIV replication in patients and lends credence to the genetic barrier hypothesis. If the low genetic barrier to resistance for this combination is the primary cause for virologic failure, it is likely that other combinations in which 2 mutations cause significant resistance to the regimen may also fail rapidly.
The authors thank Gilead Sciences for providing tenofovir for these studies and the many participants in the ESS30009 trial. We also thank Mark Underwood for preparing the plasmids for generating pHN-based mutant viruses.
1. Gallant JE, Rodriguez AE, Weinberg W, et al. Early non-response to tenofovir DF (TENOFOVIR) + abacavir (ABACAVIR) and lamivudine (lamivudine) in a randomized trial compared to efavirenz (EFV) + ABACAVIR and lamivudine: ESS30009 unplanned interim analysis. [Abstract H-840]. Presented at the 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy; Chicago, IL. September 14-17, 2003.
2. Landman R, Peytavin G, Descamps D, et al. Low genetic barrier to resistance is a possible cause of early virologic failure in once-daily regimen of abacavir, lamivudine, and tenofovir: the Tonus Study [abstract 52]. Presented at the 11th Conference on Retroviruses and Opportunistic Infections; San Francisco, CA. February 8-11, 2004.
3. Ross L, Gerondelis P, Rouse EG, et al. Impact of HIV resistance mutations, drug resistance and viral fitness on antiviral activity of tenofovir/abacavir/lamivudine in the ESS30009 study [abstract 159]. In: Programme and Abstracts of the XIII International HIV Drug Resistance Workshop (Tenerife, CI). Antivir Ther
4. Kakuda TN, Anderson PL, Becker SL. CD4 cell decline with didanosine and tenofovir and failure of triple nucleoside/nucleotide regimens may be related. AIDS
5. Miller MD, Margot N, Lu B, et al. Genotypic and phenotypic predictors of the magnitude of response to tenofovir disoproxil fumarate treatment in antiretroviral-experienced patients. J Infect Dis
6. Lanier ER, Ait-Khaled M, Scott J, et al. Antiviral efficacy of abacavir in antiretroviral therapy-experienced adults harbouring HIV-1 with specific patterns of resistance to nucleoside reverse transcriptase inhibitors. Antivir Ther
7. Miller MD, White KL, Petropoulos CJ, et al. Decreased replication capacity of HIV-1 clinical isolates containing K65R or M184V RT mutations [abstract 616]. Presented at the 10th Conference on Retroviruses and Opportunistic Infections; Boston, MA. February 10-14, 2003.
8. Selleseth DW, Talarico CL, Miller T, et al. Interactions of 1263W94 with other antiviral agents in inhibition of human cytomegalovirus replication. Antimicrob Agents Chemother
9. Daluge SM, Purifoy DJ, Savina PM, et al. 5-chloro-2′3′-deoxy-3′fluorouridine (935U83), a selective anti-human immunodeficiency virus agent with an improved metabolic and toxicological profile. Antimicrob Agents Chemother
10. Schwartz O, Henin Y, Marechal V, et al. A rapid and simple colorimetric test for the study of anti-HIV agents. AIDS Res Hum Retroviruses
11. Berenbaum MC. The expected effect of a combination of agents: the general solution. J Theor Biol
12. Tremblay CL, Giguel F, Dong H, et al. In vitro antiviral interactions among tenofovir, abacavir, lamivudine and didanosine [abstract 41]. In: Programme and Abstracts of the XIII International HIV Drug Resistance Workshop (Tenerife, CI). Antivir Ther
13. Ray A, Vela L, Olson G, et al. Lack of negative interaction between tenofovir and abacavir in human cells [abstract 2.3]. Presented at the 5th International Workshop on Clinical Pharmacology of HIV Therapy; Rome, Italy. April 1-3, 2004.
14. Hawkins T, Veikley W, St. Claire R, et al. Intracellular pharmacokinetics of tenofovir-DP and carbovir-TP in patients receiving triple nucleoside regimens [abstract 2.4]. Presented at the 5th International Workshop on Clinical Pharmacology of HIV Therapy; Rome, Italy. April 1-3, 2004.
This article has been cited
Keywords:© 2005 Lippincott Williams & Wilkins, Inc.
Tenofovir; abacavir; lamivudine; K65R; M184V; in vitro combination activity