Lack of a Metabolic and Antiviral Drug Interaction Between Tenofovir

Home , Abacavir

Antiviral Therapy 10:451–457 Lack of a metabolic and antiviral drug interaction between tenofovir, abacavir and lamivudine Adrian S Ray1*, Florence Myrick2, Jennifer E Vela1, Loren Y Olson1, Eugene J Eisenberg1, Katyna Borroto- Esoda2, Michael D Miller1 and Arnold Fridland1

1Gilead Sciences, Inc, Foster City, CA, USA 2Gilead Sciences, Inc, Durham, NC, USA

*Corresponding author: Tel: +1 650 571 7697; Fax: +1 650 522 5536; E-mail: [email protected]

Objective: An anti-HIV regimen composed of the nucleo- presence or absence of 3TC was studied in stimulated side/nucleotide reverse transcriptase inhibitors (NRTIs) PBMCs infected with the HXB2 strain of HIV-1. tenofovir (TFV) disoproxil fumarate (TDF), abacavir (ABC) Results: Levels of the active metabolites produced from and lamivudine (3TC) has performed poorly in patients. TFV and ABC after incubation with CEM or PBMCs This study evaluated the combination of TFV, ABC and showed no significant change upon introduction of the 3TC for metabolic or antiviral antagonism in vitro. other NRTI. Moreover, the pool sizes for the natural Design: Procedures were developed to evaluate the in substrates of 2′-deoxyadenosine triphosphate and 2′- vitro metabolism and antiviral activity of drug combina- deoxyguanosine triphosphate were also unchanged. In tions of TFV, ABC and 3TC in cell types relevant for HIV anti-HIV assays in PBMCs, the combination of TFV and infection. ABC was found to be additive with respect to inhibition Methods: Anabolism of combinations of TFV and ABC of HIV replication. Addition of 3TC to the combination did were studied over a 24 h period in the human T not result in synergistic or antagonistic effects. leukaemic CEM lymphoblast cell line and human primary Conclusions: The poor efficacy of the triple NRTI regimen peripheral blood mononuclear cells (PBMCs) stimulated of TDF, ABC and 3TC is probably not due to a metabolic with human interleukin-2 and phytohaemagglutinin. The drug interaction resulting in antagonism of antiviral anti-HIV activity of combinations of TFV and ABC in the activity.

Introduction

High rates of virological non-response and virological the combination and (iii) a pharmacological drug inter- failure have been observed in several clinical trials action between TDF and ABC. The lack of a systemic assessing the ability of triple nucleoside reverse tran- pharmacokinetic drug interaction between TDF and scriptase inhibitor (NRTI) therapies to suppress HIV ABC in plasma has already been established [4]. infection. Study ESS30009 was conducted in 194 treat- The oral prodrug TDF is converted to TFV upon ment-naive patients and compared the efficacy of teno- absorption. Within cells, both ABC and TFV require fovir (TFV) disoproxil fumarate (TDF)/abacavir intracellular phosphorylation to their nucleoside (ABC)/lamivudine (3TC) to efavirenz/ABC/3TC, all in triphosphate analogue forms in order to have antiviral once-daily regimens. In an unplanned interim analysis, activity. The intracellular anabolism of ABC to its active a 49% virological failure rate was observed at 12 2′-deoxyguanosine triphosphate (dGTP) analogue, weeks in the triple NRTI arm versus 5% in the carbovir triphosphate (CBV-TP) is complex. ABC is efavirenz-containing arm [1]. Other studies with the first phosphorylated by adenosine monophosphate same combination of TDF/ABC/3TC also found poor phosphotransferase to ABC monophosphate, which is response and a high incidence of viral resistance muta- then deaminated to carbovir monophosphate (CBV- tions [2,3]. Since TDF and ABC have each been used MP) and then phosphorylated to CBV-TP via two successfully in regimens with 3TC, a problem between cellular kinases (Figure 1A) [5]. TFV follows a different TDF and ABC has been surmised. These results have led anabolic pathway to the active 2′-deoxyadenosine to the proposal of different mechanisms for the poor triphosphate (dATP) analogue TFV diphosphate (TFV- efficacy of triple NRTI regimens containing TDF and DP). The activation pathway involves the formation of ABC, including (i) intrinsic poor potency of all triple a single key intermediate, TFV-MP, by the activity of NRTI regimens, (ii) a low barrier for viral resistance to adenylate kinase 2 followed by further phosphorylation

Figure 1. Activation pathways for ABC and TFV

A O HN HN N N N H2N N N N H2N N AMP O HO –O PO Carbovir-TP O– AMP Cellular Cellular ABC phosphotransferase deaminase Carbovir-MP kinases

B NH2 N N

N O N –O PO Tenofovir-DP O– Adenylate Cellular kinase 2 kinases TFV

(A) Anabolism of ABC includes deamination as well as phosphorylation [5]. (B) Phosphorylation pathway for TFV based on that determined for the acyclic nucleotide analogue adefovir [21]. Carbovir-MP, carbovir monophosphate; carbovir-TP, carbovir triphosphate; tenofovir-DP; tenofovir diphosphate. ABC, abacavir; TFV, tenofovir.

to the active metabolite TFV-DP (Figure 1B) [6]. TFV- by GlaxoSmithKline (Research Triangle Park, NC, DP and CBV-TP compete with the intracellular dATP USA). All other chemicals used were the highest grade and dGTP pools, respectively, for incorporation by the available from Sigma-Aldrich, Inc (St Louis, MO, USA). viral RT. Due to their lack of 3′-hydroxyl groups, incorporation of these nucleotide analogues causes Cell culture chain termination of viral reverse transcripts and inhi- Human T leukaemic CCRF-CEM cells were obtained bition of viral replication. from the American Type Culture Collection (Manassas, In this article, we further examine the potential for a VA, USA) and were maintained in RPMI 1640 supple- drug interaction between TFV and ABC by studying mented with 10% heat-inactivated fetal bovine serum the in vitro intracellular pharmacology and anti-HIV (FBS), 100 units/ml penicillin G and 100 µg/ml strepto- activity of the two agents incubated either alone or in mycin sulphate. Stanford Blood Center (Palo Alto, CA, combination in cell types relevant for HIV infection. USA) and the American Red Cross (Durham, NC, USA) supplied buffy coats prepared from the blood of Materials and methods healthy donors. Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats by density Chemicals gradient centrifugation on Ficoll-Paque Plus using stan- Cell culture supplies were purchased from Irvine dard techniques as recommended by the manufacturer Scientific (Santa Ana, CA, USA). TFV and TFV-DP as (Amersham Pharmacia Biotech, Inc, Piscataway, NJ, well as adefovir-DP (used as an internal standard) were USA). Two additional washes in RBC lysis buffer synthesized by Gilead Sciences, Inc (Foster City, CA, (eBioscience, San Diego, CA, USA) were performed for USA). Stable isotope-labelled 13C15N dATP and dGTP PBMC preparations used in assays for triphosphate were purchased from Spectra Stable Isotopes evaluations. PBMC pellets were resuspended in 1640 (Columbia, MD, USA). TFV-MP was synthesized by Dr RPMI medium containing 10% heat-inactivated FBS, Ivan Rosenberg of the Institute of Organic Chemistry 2 mM L-glutamine and penicillin/streptomycin. Cells and Biochemistry, Czech Academy of Sciences, Prague, were stimulated with 20 U/ml of human interleukin-2 Czech Republic. ABC, 3TC and CBV-TP were provided (IL-2; Roche Diagnostics Corp, Indianapolis, IN, USA)

and 1–5 µg/ml of phytohaemagglutinin from Phaseolus Antiviral synergy vulgaris (PHA-P; Sigma-Aldrich, Inc) for 48–72 h at TFV and ABC were tested in combination for antiviral 37°C and subsequently cultured in the same medium activity against the HXB2 laboratory strain of HIV-1 in without PHA-P. PBMCs. Stimulated PBMCs were infected with virus at a multiplicity of infection of 0.001 for 4 h, washed Intracellular metabolism once with complete media and plated in 96-well plates TFV or ABC either alone or in combination at concen- containing various concentrations of test compound. trations of 10–100 µM were added to CEM cells The final concentration of cells was 1 million PBMC seeded at 0.2 million/ml or PBMCs seeded at per well. Compounds were plated in a checkerboard 0.5 million/ml. At 2 and 24 h, aliquots of cells taken fashion with the starting concentration for each for analyses were separated from the media and extra- compound fixed at three- to fourfold above the EC50 cellular drug by spinning through oil [7]. Isolated cells value for the specific compound. Subsequent twofold were resuspended in 70% methanol in order to extract serial dilutions were then performed across the plate. intracellular metabolites. After drying by vacuum, Five plates for each drug combination were tested. The ° samples were resuspended in an injection buffer plates were incubated for 4 days at 37 C in 5% CO2 containing 10 mM tetrabutyl ammonium hydroxide and the level of virus replication monitored by evalua- (TBAH) and 10 mM ammonium phosphate pH 7.0 at tion of HIV-1 p24 antigen in the cell culture super- concentrations of 1–5 million cells per 20 µl, and used natant of each well (Coulter HIV-1 p24 antigen assay). for analysis of intracellular anabolites. Combinations of 3TC and 3TC were included as Transient ion-pairing high performance liquid chro- controls for additivity. Combinations of TFV with 3TC matography (HPLC) coupled to positive ion electro- and ABC with 3TC were also evaluated as positive spray tandem mass spectrometry (LC/MS/MS) was controls as these compounds have demonstrated good used to analyse TFV and ABC in their phosphorylated clinical activity when used in combination. forms for uptake and egress studies. Earlier methods Results from the antiviral assays were analysed using for nucleoside triphosphate analysis using TBAH as an the isobologram method [9] and the MacSynergy ion pair [8] were modified to include an acetonitrile program [10] to determine potential synergistic and/or step gradient in order to attain sufficient retention of antagonistic activities. The isobologram evaluation is TFV and TFV-MP on an Xterra MS, C18, 3.5 µm, based on the Lowe additivity model. Dose response 1.0×100 mm reverse phase column (Waters Corp, curves were generated for each drug alone and in

Milford, MA, USA). Mobile phases included 0.25 mM combination and used to determine EC50 values for TBAH, 4 mM ammonium phosphate pH 6.5 and each drug alone or in the presence of the fixed concen- acetonitrile concentrations of 6 and 20%. A flow rate tration of the second drug. Fractional inhibitory of 50 µl/min was maintained using HPLC with an concentrations (FICs) were calculated by dividing the

Agilent 1100 chromatographic system (Agilent EC50 of drug 1 with a fixed overlay of drug 2 by the

Technologies, Palo Alto, CA, USA). Parent/daughter EC50 of drug 1 alone (the x coordinate). The y coordi- mass transitions of 274.2/162.4, 354.0/162.4, nate is the fixed concentration of drug 2 divided by the

448.3/176.4, 488.0/135.3 and 434.0/162.4 m/z for EC50 of drug 2 alone. These points are plotted on a TFV, TFV-MP, TFV-DP, CBV-TP and adefovir-DP graph to generate the isobologram. On this same (internal standard), respectively, were monitored on an graph, a line representing additivity is included [coor- API-4000 triple quadrupole mass spectrometer dinates (1,0) to (0,1)]. Data points that are above the purchased from Applied Biosystems/MDS Sciex (Foster additivity line represent antagonism between the City, CA, USA). Typical lower limits of quantitation of compounds whereas data points below the additivity 20–200 fmols on column (0.01–0.04 pmols/million line represent synergy between the compounds. The cells in a 20 µl injection of lysate from 1–5 million cells) intensity of the interaction can be quantitated and the were achieved. Stable isotope methods for the measure- statistical difference from dose-wise additivity tested by ment of intracellular dATP and dGTP will be described a 1-tailed t-test [11]. in detail elsewhere (manuscript in preparation). Briefly, The MacSynergy program calculates a theoretical standard curves in cellular matrix were performed additive value for each drug combination based on the utilizing stable isotope-labelled dATP and dGTP (mass values generated by the drugs alone using an indepen- transition of 506.7/145.6 and 522.6/161.8 m/z) while, dent effects model (Bliss independence). The theoretical concurrently, the amount of dATP and dGTP (mass additive values are subtracted from the experimental transition of 492.1/136.2 and 507.7/151.6 m/z) in the values generated by each drug combination to give a cellular sample were monitored. The LC/MS/MS value of synergy (positive value) or antagonism (nega- methods described above for TFV-DP and CBV-TP tive value). These synergy and/or antagonism values detection were used for these analyses. are plotted on a three-dimensional graph with their

Antiviral Therapy 10:3 453 AS Ray et al.

corresponding drug combinations. Areas of the graph unaffected by the presence of TFV (Table 1). To test if below zero indicate antagonism whereas areas above supra-pharmacological concentrations of the drugs zero indicate synergy. A synergy volume is calculated could affect their metabolism, experiments were by adding all of the synergy values (positive values) for repeated in CEM cells incubated with 100 µM TFV each drug combination. Likewise, all of the antago- and ABC. While levels of intracellular metabolites for nistic values (negative values) are added to give an both drugs were increased by approximately 10-fold, antagonistic volume. These synergy and antagonism no difference in the accumulation of TFV or ABC volumes are then statistically evaluated using the 95% anabolites were noted upon co-administration (data confidence level and are expressed in µM2%. Values not shown). To verify that it is possible to elicit a between –25 µM2% and +25 µM2% are considered nucleoside drug interaction at the level of phosphory- additive, values below –25 µM2% represent antago- lation in CEM-CCRF cells, studies were done with nism and values above 25 µM2% represent synergy. 3TC and zalcitabine (ddC) incubated at 10 µM. Consistent with previous reports [12], levels of ddC-TP Results were reduced fivefold (from 102 to 22.6 pmol/million cells) upon the addition of 10 µM 3TC to the incuba- Metabolism of TFV and ABC in CEM cells tion media (data not shown). To address whether ABC and TFV interacted with one another’s phosphorylation, initial experiments were Effect of TFV and ABC on 2′ deoxynucleotide pool done examining the phosphorylation of TFV and ABC in sizes the T-cellline CEM-CCRF.The cells were incubated with After phosphorylation to their respective triphosphate TFV (10 µM) or ABC (10 µM) and the level of phospho- analogues, TFV and ABC compete with the dATP or rylated anabolites formed were determined using dGTP deoxynucleotide pools for incorporation by LC/MS/MS analysis after 24 h of incubation. CEM cells HIV-1 reverse transcriptase. Alterations in the accumulated levels of TFV-MP and TFV-DP of 0.53 and competing pool sizes could negatively affect the 1.49 pmol/million cells, respectively,after 24 h (Table1). antiviral activity of TFV and ABC. The observation of The addition of 10 µM ABC did not markedly affect the the inhibition of an enzyme involved in the regulation formation of TFV-DP (1.44 pmols/million cells). The of nucleotide pools (purine nucleoside phosphorylase) conversion of ABC to CBV-TP was found to be less effi- by phosphorylated metabolites of TFV [7] has led cient than TFV, reaching levels of 0.12 pmols/million Kakuda and colleagues to hypothesis that TFV may after 24 h. The anabolism of ABC at 24 h was alter nucleotide pools [13]. To test if TFV or ABC can alter dATP or dGTP pools, 10 µM incubations were carried out either alone or in combination with CEM- CCRF cells and levels of dATP and dGTP were deter- Table 1. In vitro metabolism of 10 µM TFV and 30 µM ABC mined after 24 h by LC/MS/MS. CEM cells were either alone or in combination and their effects on dATP and dGTP pool sizes chosen for these studies because it has been shown that they are highly sensitive to pool size alterations by Intracellular concentration* inhibitors of purine nucleoside phosphorylase [14]. Cell type Nucleotide TFV alone ABC alone TFV + ABC Results in Table 1 show that neither ABC nor TFV, either alone or in combination, significantly reduced CEM TFV 1.42 ±0.41 1.59 ±0.01 the competing deoxynucleotide pool sizes from the TFV-MP 0.53 ±0.12 0.47 ±0.08 42.7 ±10.7 and 23.7 ±10.9 pmols/million cell levels ± ± TFV-DP 1.49 0.07 1.44 0.20 measured in untreated cells for dATP and dGTP, ± ± CBV-TP 0.12 0.07 0.18 0.09 respectively. dATP† 40.3 ±7.6 40.3 ±4.8 41.7 ±12.2 † ± ± ± dGTP 20.5 7.3 22.8 2.6 23.8 14.6 Metabolism of TFV and ABC in stimulated PBMCs PBMC TFV 1.11 ±0.47 0.81 ±0.05 TFV-MP 0.15 ±0.06 0.16 ±0.10 We extended these combination studies of TFV and TFV-DP 0.21 ±0.06 0.22 ±0.11 ABC in the cell type most physiologically relevant for CBV-TP 0.11 ±0.03 0.11 ±0.01 HIV therapy, human PBMCs. PBMCs from two separate donors were stimulated with IL-2 and PHA. *Concentrations expressed in pmol/106 cells for all measurements after 24 h of Twenty-four h incubations with 10 µM TFV and drug incubation. Values for CEM cells are the mean ±SD of n=4 experiments. Values for stimulated human PBMCs are the mean ±SD of n=4 experiments done 30 µM ABC resulted in the intracellular activation of in cells from two separate donors. †Intracellular concentrations of dATP and each analogue. The results summarized in Table 1 dGTP in untreated CEM cells were 42.7 ±10.7 and 23.7 ±10.9 pmols/million cells, respectively. dATP, 2′-deoxyadenosine triphosphate; dGTP, 2′-deoxyguanosine demonstrate that there was no difference in the levels triphosphate; PBMC, peripheral blood mononuclear cell; ABC, abacavir; CBV, of TFV-DP or CBV-TP formed after 24 h whether the carbovir; TFV, tenofovir; -MP, monophosphate; -DP, diphosphate; -TP, triphosphate. drugs were incubated alone or together. Similar to

Table 2. Summary of MacSynergy analysis

Drug 1 Drug 2 3TC overlay Synergy volume, µM2% Antagonism volume, µM2% Net effect*, µM2%

ABC 3TC None 45.1 –14.3 30.8 (minor synergy) TFV 3TC None 33.8 0 33.8 (minor synergy) ABC TFV None 6.1 –20.6 –14.5 (additive) ABC TFV 25 nM 1.7 –0.5 1.2 (additive) ABC TFV 50 nM 2.8 –9.8 –7 (additive) ABC TFV 150 nM 10.3 0 10.3 (additive) 3TC 3TC None 19.1 –1.8 17.3 (additive)

*Net effect calculated as the sum of synergy and antagonism volumes. Values below –25 µM2% are antagonistic and values greater than 25 µM2% are synergistic; values between –25 µM2% and +25 µM2% are additive. 3TC, lamivudine; ABC, abacavir; TFV, tenofovir.

results obtained in CEM cells, quantitation of data for Figure 2. Isobologram for the inhibition of HIV replication by TFV and TFV-MP also showed that there were no the combination of ABC and TFV significant differences beyond standard experimental error (Table 1). 1.4 Antiviral interaction between TFV and ABC in the presence or absence of 3TC 1.2 Experimental Additive To expand upon these intracellular metabolism studies, 1 TFV and ABC were analysed for synergistic or antagonistic anti-HIV activity in PBMCs infected with the 0.8 HXB2 strain of HIV-1. Using MacSynergy analysis, the

combination of TFV and ABC was found to be additive FIC (TFV) 0.6 with regards to inhibition of HIV-1 replication (Table 2). The combination of TFV and 3TC and the 0.4 combination of ABC and 3TC showed minor synergy 0.2 with net synergy volumes >25 µM2%. Triple combinations were performed by addition of 3TC at concentra- 0 tions of 25 nM, 50 nM and 150 nM to graded 0 0.2 0.4 0.6 0.8 1 1.2 1.4 concentrations of TFV and ABC. The addition of 3TC at any of these concentrations had no effect on the FIC (ABC) observed additive anti-HIV activity of TFV and ABC. Higher concentrations of 3TC in combination with Theoretical additivity is represented by the straight line which intercepts both TFV and ABC resulted in complete suppression of HIV FIC at axes at a value of 1. Error bars show the standard error of the two FICs plotted. FIC, fractional inhibitory concentrations; ABC, abacavir; TFV, tenofovir. growth and thus synergy calculations were not attain- able (data not shown). As expected, the combination of 3TC and 3TC showed additive anti-HIV activity. The combination anti-HIV assays were also analysed for synergy or antagonism by the isobologram Table 3. Summary of isobologram synergy analysis method. A representative isobologram for the combi- Average deviation from additivity nation of TFV and ABC is shown in Figure 2. At each † † drug concentration, the fractional inhibitory concen- Drug combination Mean* SEM P value Net effect trations were below the additivity line indicating a lack ABC + TFV –0.2137 0.1159 0.057 Additive of antagonism and possible synergy. The other two- TFV + 3TC –0.2206 0.0918 0.027 Synergistic drug combinations and the three-drug combination ABC + 3TC –0.0417 0.0844 0.323 Additive showed similar results with negative mean deviations ABC + TFV + 3TC –0.0531 0.2222 0.410 Additive from additivity (Table 3). Statistical analysis indicated that deviation from additivity was significant only for *Mean deviation from theoretical additivity line for all drug concentrations. † the two-drug combination of TFV and 3TC. No combi- Negative mean deviation indicates additivity. For P values <0.05, synergy is indicated. SEM, standard error of the mean; 3TC, lamivudine; ABC, abacavir; nations showed antagonistic anti-HIV activity. TFV, tenofovir.

Antiviral Therapy 10:3 455 AS Ray et al.

Conclusions failure in some patients. Further support for this hypothesis comes from the clinical observations that Triple NRTI therapies have been clinically explored patients with high viral loads are more likely to fail on because they could theoretically have several advan- these regimens with a step-wise development of resis- tages over non-NRTI (NNRTI)- and protease inhibitor tance [20]. Better understanding of the underlying (PI)-containing regimens. Triple NRTI treatments mechanism responsible for the poor efficacy of these could allow for infrequent dosing with a small number regimens may help guide the development of more of pills, avoidance of the unwanted side effects and effective combination regimens. drug interactions associated with the other classes of HIV therapy, and, in first-line treatment, leave options Acknowledgements open for later administration of NNRTI- or PI- containing regimens. Unfortunately, the promising The drugs 3TC, ABC and CBV-TP were provided by potential for triple NRTI therapies has not been real- GlaxoSmithKline. ized and poor clinical outcomes have been observed in several studies [15]. Conflicts of interest The goal of the experiments described in this article was to determine if there is an intracellular drug inter- All authors are employees of Gilead Sciences, Inc, the action between ABC and TFV that could potentially marketer of tenofovir disoproxil fumarate (Viread). explain the poor performance of triple NRTI therapies containing these two agents. In order to accomplish References this, studies were conducted in cell types relevant for 1. Gallant JE, Rodriguez AE, Wieinberg W, Young B, Berger HIV infection. The results described in this article D, Lim ML, Liao Q, Ross L, Johnson J, Shaefer M & Team found no evidence for an intracellular drug interaction ES. Early non-response to tenofovir-DF (TDF) + abacavir (ABC) and lamivudine (3TC) in a randomized trial between TFV and ABC using both a direct intracellular compared to efavirenz (EFV) + ABC and 3TC: ESS30009 drug metabolism analysis and antiviral drug combina- unplanned interim analysis. 43rd International Conference on Antimicrobial Agents & Chemotherapy. 14–17 tion analysis in PBMCs. The recently reported September 2003, Chicago, IL, USA. Abstract H-17722a. measurements of intracellular CBV-TP and TFV-DP in 2. Farthing C, Khanlou H, Yeh V & Harris G. Early virologic PBMCs from HIV-infected patients undergoing therapy failure in a pilot study evaluating the efficacy of once daily with these two drugs in combination have led to similar abacavir (ABC), lamivudine (3TC) and tenofovir DF (TDF) in treatment naive HIV-infected patients. 2nd International conclusions [16]. These results are consistent with the AIDS Society Conference on HIV Pathogenesis & distinct anabolic pathways presented in Figure 1 for Treatment. 13–16 July 2003, Paris, France. Abstract S195. these two NRTIs. 3. Hoogewerf M, Regez RM, Schouten WE, Weigel HM, Frissen PH & Brinkman K. Change to abacavir-lamivudine- The result of these in vitro studies, as well as data tenofovir combination treatment in patients with HIV-1 reported elsewhere showing a lack of detectable intra- who had complete virological suppression. Lancet 2003; 362:1979–1980. cellular interaction in vivo [16] and no systemic plasma 4. Kearney BP, Isaacson E, Sayre J, Ebrahimi R & Cheng AK. drug interaction in vivo [4], suggest that the poor effi- The pharmacokinetics of abacavir, a purine nucleoside cacy of triple NRTI combinations including TDF and analogue, are not affected by tenofovir DF. 43rd Annual Interscience Conference on Antimicrobial Agents & ABC cannot be attributed to a drug–drug interaction Chemotherapy. 14–17 September 2003, Chicago, IL, USA. between TDF and ABC. Furthermore, the results Abstract A-1615. presented here show that neither ABC, TFV or their 5. Faletto MB, Miller WH, Garvey EP, St Clair MH, Daluge SM & Good SS. Unique intracellular activation of the combination altered the intracellular pool sizes of potent anti-human immunodeficiency virus agent 1592U89. dATP or dGTP, illustrating that perturbations in the Antimicrobial Agents & Chemotherapy 1997; natural dNTP pools is an unlikely explanation for the 41:1099–1107. poor efficacy observed. 6. Krejcova R, Horska K, Votruba I & Holy A. Phosphorylation of purine (phosphonomethoxy)alkyl deriv- The poor performance of other triple NRTI regi- atives by mitochondrial AMP kinase (AK2 type) from mens not including TDF and ABC together argues L1210 cells. Collection of Czechoslovak Chemical Contributions 2000; 65:1653–1668. against a specific mechanism for suboptimal response. 7. Ray AS, Olson L & Fridland A. Role of purine nucleoside Combinations of didanosine/stavudine/ABC and phosphorylase in interactions between 2′,3′-dideoxyinosine TDF/3TC/didanosine have also performed poorly and allopurinol, ganciclovir, or tenofovir. Antimicrobial Agents & Chemotherapy 2004; 48:1089–1095. [17,18]. The early termination of the zidovudine/ABC/ 8. Claire RL 3rd. Positive ion electrospray ionization tandem 3TC arm in the ACTG 5095 study also raises concerns mass spectrometry coupled to ion-pairing high-performance about this triple NRTI combination [19]. These results liquid chromatography with a phosphate buffer for the quantitative analysis of intracellular nucleotides. Rapid lend indirect evidence for the hypothesis that there is Communications Mass Spectrometry 2000; 14:1625–1634. weaker potency in triple NRTI regimens, which can 9. Elion GB, Singer S & Hitchings GH. Antagonists of nucleic lead to the development of resistance and treatment acid derivatives. VIII. Synergism in combinations of

biochemically related antimetabolites. Journal of Biological Pharmacology of HIV Therapy. 1–3 April 2004, Rome, Chemistry 1954; 208:477–488. Italy. Abstract 5. 10. Prichard MN & Shipman C Jr. Analysis of combinations of 17. Gertstoft J, Kirk O, Obel N, Pederson C, Mathiesen L, antiviral drugs and design of effective multidrug therapies. Nielsen H, Katzenstein TL & Lundgren JD. Low efficacy Antiviral Therapy 1996; 1:9–20. and high frequency of adverse events in a randomized trial of the triple nucleoside regimen abacavir, stavudine and 11. Selleseth DW, Talarico CL, Miller T, Lutz MW, Biron KK didanosine. AIDS 2003; 17:2045–2052. & Harvey RJ. Interactions of 1263W94 with other antiviral agents in inhibition of human cytomegalovirus 18. Winston A, Pozniak A, Mandalia S, Gazzard B & Nelson replication. Antimicrobial Agents & Chemotherapy 2003; M. Which nucleoside and nucleotide backbone combina- 47:1468–1471. tions select for the K65R mutation in HIV-1 reverse transcriptase. 12th International HIV Drug Resistance 12. Hoggard PG, Sales SD, Kewn S, Sunderland D, Khoo SH, Workshop. 10–14 June 2003, Los Cabos, Mexico. Abstract Hart CA & Back DJ. Correlation between intracellular 137. pharmacological activation of nucleoside analogues and HIV suppression in vitro. Antiviral Chemistry & 19. Gulick RM, Ribaudo HJ, Shikuma CM, Lustgarten S, Squires Chemotherapy 2000; 11:353–358. KE, Meyer WA 3rd, Acosta EP, Schackman BR, Pilcher CD, Murphy RL, Maher WE, Witt MD, Reichman RC, Snyder S, 13. Kakuda TN, Anderson PL & Becker SL. CD4 cell decline Klingman KL & Kuritzkes DR. Triple-nucleoside regimens with didanosine and tenofovir and failure of triple nucleo- versus efavirenz-containing regimens for the initial treat- side/nucleotide regimens may be related. AIDS 2004; ment of HIV-1 infection. New England Journal of Medicine 18:2442–2444. 2004; 350:1850–1861. 14. Bantia S, Ananth SL, Parker CD, Horn LL & Upshaw R. 20. Delauney C, Descamps D, Landman R, Peytavin G, Falndre Mechanism of inhibition of T-acute lymphoblastic leukemia P, Trylesinski A, Collin G, Benalycherif A, Monchecourt F, cells by PNP inhibitor – BCX-1777. International Brun-Vézinet F & Group TTS. Dynamic of selection of the Immunopharmacology 2003; 3:879–887. K65R and M184V/I mutations in patients enrolled in 15. Nicholson M. No nukes? AIDS Treatment Update 2004; Tonus Trial. Antiviral Therapy 2004; 9:S171. 133:6–9. 21. Robbins BL, Greenhaw J, Connelly MC & Fridland A. 16. Hawkins T, Veikley W, StClaire R, Hey A, Guyer B & Metabolic pathways for activation of the antiviral agent 9- Kearney BP. Intracellular pharmacokinetics of tenofovir-DP (2-phosphonylmethoxyethyl)adenine in human lymphoid and carbovir-TP in patients receiving triple nucleoside regi- cells. Antimicrobial Agents & Chemotherapy 1995; mens. 5th International Workshop on Clinical 39:2304–2308.

Received 8 December 2004, accepted 20 March 2005

Antiviral Therapy 10:3 457