Influence of Prior Exposure to Zidovudine on Stavudine Phosphorylation In Vivo and Ex Vivo

Abstract

Intracellular phosphorylation of stavudine (d4T) and zidovudine (ZDV) was investigated in peripheral blood mononuclear cells (PBMCs) isolated from ZDV-naive and ZDV-experienced human immunodeficiency virus (HIV)-positive patients. An in vivo study measured the amount of d4T triphosphate (d4TTP), while an ex vivo study assessed the capacity of cells to phosphorylate added d4T. Endogenous dTTP was also measured. d4TTP and dTTP were determined in vivo using a reverse transcriptase chain termination assay. In ex vivo studies, d4T (1 μM) was incubated in resting and phytohemagglutinin-stimulated (10 μg ml⁻¹; 72 h) PBMCs for 24 h. After washing and methanol extraction, radiolabeled anabolites were detected by high-performance liquid chromatography. d4TTP reached its highest level 2 to 4 h after dosing (0.21 ± 0.14 pmol/10⁶ cells; n = 27 [mean ± standard deviation]). Comparison of ZDV-naive and ZDV-experienced individuals showed no significant difference in levels of d4TTP (ZDV naive, 0.23 ± 0.17 pmol/10⁶ cells [n = 7] versus ZDV experienced, 0.20 ± 0.14 pmol/10⁶ cells [n = 20]; P = 0.473) or the d4TTP/dTTP ratio (0.14 ± 0.12 [n = 7] and 0.10 ± 0.08 [n = 20], respectively; p = 0.391). Ex vivo data demonstrated no significant difference in the formation of d4TTP or total d4T phosphates in naive and experienced patients (0.086 ± 0.055 pmol/10⁶ cells in ZDV-naive patients [n = 17] versus 0.081 ± 0.038 pmol/10⁶ cells in ZDV-experienced patients [n = 22]; P = 0.767). The ability of HIV-infected patients to phosphorylate d4T in vivo and ex vivo was unchanged with increasing exposure to ZDV.

The sequence in which human immunodeficiency virus (HIV) nucleoside analogues such as zidovudine (ZDV) and stavudine (d4T) should be used has been the subject of much debate in recent years. Preliminary observations from in vitro studies and some small clinical studies have suggested that the intracellular phosphorylation of these compounds may vary with duration of therapy or prior use of another nucleoside analogue [1, 4; J. P. Sommadossi, X. Zhou, J. Moore, D. R. Havlir, G. Friedland, C. Tierney, L. Smeaton, L. Fox, D. Richman, and R. Pollard, 5th Conf. Retrovir. Opportunistic Infect., 1998).

The intracellular phosphorylation of these drugs to their active triphosphates is essential for anti-HIV activity. The ratio of drug triphosphates to endogenous deoxynucleoside triphosphates (dNTPs) most accurately reflects the direct competition between drug triphosphate and dNTP for HIV reverse transcriptase (RT) (6, 7, 21). Treatment failure is multifactorial. In addition to viral resistance and lack of adherence, it has been postulated that decreased intracellular phosphorylation may contribute to the loss of therapeutic benefits of drugs seen over time (4, 10). A number of factors have been implicated, including previous drug experience and the disease stage. HIV infection may alter the capacity of peripheral blood mononuclear cells (PBMCs) to phosphorylate drugs by reducing thymidine kinase activity (12, 13). Prolonged use of one nucleoside or prior exposure to other nucleoside analogues may modulate the activity of phosphorylating enzymes. Thymidine kinase has a 600-fold-lower affinity for d4T than for thymidine and ZDV and represents the rate-limiting step for d4T phosphorylation (11). Thus, changes in d4T phosphorylation in ZDV-experienced patients would be the most sensitive indicator of ZDV-induced down-regulation of thymidine kinase, if this mechanism was important. The ALTIPHAR study reported that patients exposed to ZDV had decreased formation of d4T and lamivudine (3TC) triphosphates compared to those who were ZDV naive (Sommadossi, et al., 5th Conf. Retrovir. Opportunistic Infect.).

We investigated the intracellular phosphorylation of d4T in PBMCs from ZDV-naive and ZDV-experienced patients. An in vivo study measured the amount of d4T triphosphate (d4TTP) and dTTP formed, and an ex vivo study examined the capacity of PBMCs to phosphorylate added d4T.

MATERIALS AND METHODS

Materials.

Lymphoprep was purchased from Nycomed Pharma AS, Oslo, Norway. d4T and [methyl-³H]d4T were gifts from Bristol-Myers Squibb, Wallingford, Conn. ZDV was donated by the MRC AIDS Research Project. [methyl-³H]ZDV, [³H]dTTP (58 mCi mmol⁻¹), and [³H]dATP (18 mCi mmol⁻¹) were purchased from Moravek Biochemicals, Brea, Calif. d4TTP was synthesized by Sierra Biotech) Poly(rA) · oligo(dTP)12-18 was purchased from Pharmacia Biotech. Sequenase version 2.0 DNA polymerase and oligonucleotides (19) were purchased from United States Biochemical (Cleveland, Ohio). HIV RT was obtained from Amersham, Little Chalfont, United Kingdom. All other drugs and chemicals were purchased from Sigma Chemical Company Ltd. ³H-nucleoside analogues were of ≥97% purity.

Subjects.

This study was carried out between May and December 1998. Healthy volunteers and HIV-positive patients (aged 20 to 60 years; CD4 counts, 12 × 10⁶ to 640 × 10⁶/liter) were recruited. All patients were receiving antiretroviral therapy that included d4T; patients were excluded if their current regimens included either ZDV or hydroxyurea. Approval for the study was obtained (from the ethics committees of the Royal Liverpool University Hospital and North Manchester Health Authority), and each participant provided written informed consent.

Blood sampling.

A single blood sample (25 ml) was obtained by venipuncture at different times after dosing (range, 0 to 12 h).

Samples from 36 patients were designated for the in vivo study. Of these patients, 13 were ZDV naive and 23 had previously received ZDV for 1 to 25 months (median, 13 months). The time since discontinuing ZDV was 2 to 45 months (median, 24 months). PBMCs were isolated as described previously (20). Following extraction with 60% methanol and perchloric acid (20), dTTP and d4TTP were measured.

Samples from 39 patients were designated for the ex vivo study. After isolation, the PBMCs were incubated with radiolabeled d4T to determine intracellular phosphate anabolite concentrations (see below). Of these patients, 16 were ZDV naive and 23 had previously received ZDV for 1 to 84 months (median, 8 months). The time since discontinuing ZDV was 0 to 28 months (median, 16 months).

PBMCs were also isolated from 45 healthy volunteers and 45 HIV patients to investigate the effect of HIV infection on the capacity to phosphorylate radiolabeled thymidine and its analogues, ZDV and d4T, in resting and stimulated PBMCs.

dTTP assay.

Intracellular dTTP pools were measured by an enzymatic assay adapted from that of Sherman and Fyfe (19) as previously described (20).

Determination of the d4TTP standard inhibition curve.

In vivo d4TTP concentrations were determined by using an RT assay similar to that previously described for the quantification of ZDV triphosphate (ZDVTP) (16, 17). A total volume of 50 μl of the reaction mixture contained the following final concentrations: 50 mM Tris HCl (pH 7.8 to 8.5), 1 mM dithiothreitol, 6 mM MgCl₂, 30 mM KCl, 0.1% Triton X-100, 2 mM CuSO₄, 0.025% bovine serum albumin; 0.5 mM EDTA, 4 pmol (0.3 μCi) of [³H]dTTP, 0.002 μg of poly(rA) · oligo(dT)_12-18, 5 μl of cell extract (containing a known amount of dTTP), and either 0, 0.025, 0.05, 0.075, or 0.1 pmol of d4TTP. d4TTP in PBMCs from HIV patients was determined from 10-μl extracts.

Initially, d4TTP, [³H]dTTP, cell extract, Tris-HCl, and KCl were incubated with CuSO₄ at room temperature for 20 min to inactivate any RNase present in the extract. Bovine serum albumin and EDTA were then added, and the mixture was incubated for a further 15 min to remove excess CuSO₄. The template primer was included prior to initiation of the reaction by the addition of 0.4 U of HIV RT, which was then incubated in a water bath (37°C) for 60 min. The reaction mixtures (20 μl) were spotted onto filter paper (presoaked with 5% trichloroacetic acid plus 1% sodium pyrophosphate solution). Following heat drying, the filter papers were washed three times with the same cold solution (10 ml; 5 min) and finished by washing them twice with 95% ethanol (vol/vol; 3 ml). After being heat dried, the filter papers were transferred to scintillation vials and the radioactivity was counted.

Incubation of PBMCs with ³H-nucleoside analogues ex vivo.

Freshly isolated PBMCs were stimulated with phytohemagglutinin (PHA) (10 μg ml⁻¹) and cultured for 72 h in RPMI medium supplemented with 10% fetal calf serum (4 ml) and l-glutamine (2 mM) at 37°C in an incubator with humidified 5% CO₂. Freshly isolated nondividing resting PBMCs and PHA-stimulated PBMCs were then incubated with d4T (1 μM; 5 μCi) for 24 h at 37°C in an incubator with humidified 5% CO₂. Experiments were performed in triplicate.

Resting and PHA-stimulated PBMCs were also incubated with either ZDV (0.1 μM; 5 μCi) or thymidine (1 μM; 5 μCi) for 24 h as described above.

Following incubation, the cell suspensions were washed in phosphate-buffered saline and extracted in 60% methanol. The methanol extracts were evaporated to dryness and stored at −20°C until they were analyzed by high-performance liquid chromatography (HPLC).

The coefficient of variation for repeated analysis of ZDV and d4T phosphates extracted from the same sample was less than 10%. The extent of and variability in the phosphorylation of ZDV and d4T were similar to those in previous reports (2, 3, 18). In resting cells, total phosphates of ZDV, d4T, and thymidine were always detectable following a 24-h incubation. The limit of quantification of individual phosphate anabolites for d4T and thymidine was 0.003 pmol/10⁶ cells, and for ZDV the limit was 0.0003 pmol/10⁶ cells, corresponding to a peak height of three times the baseline.

HPLC analysis.

Cell extracts were reconstituted in double-distilled H₂O and separated by HPLC on an anion-exchange column (Partisil 10-SAX; 25 cm by 4.6 mm) as previously described (2, 3). Samples were eluted with increasing buffer concentrations, and phosphate metabolites were identified by cochromatography with authentic standards or occasionally by their elution order as previously established (2, 3). The eluent was collected in insert vials at 30-s collection times. Scintillant (4 ml) was added to the collected fractions, which were then counted in a liquid scintillation counter.

Statistical analysis.

Data from resting and PHA-stimulated PBMCs were standardized to picomoles per million cells. In resting cells, the results were calculated using cell counts taken at the conclusion of the drug incubations. However, since PHA causes clumping, which prevents hemocytometer cell counting at the end of the incubations, cell counts prior to mitogen stimulation were used in these incubations. d4TTP data obtained in vivo were analyzed by an unpaired t test. d4TTP data obtained ex vivo were analyzed by the Mann-Whitney U test.

Multiple linear regression followed by analysis of variance was performed to assess the dependence of d4TTP on time on and time since ZDV therapy both in vivo and ex vivo. This analysis was also used to investigate the dependence of total d4T phosphates on time on and time since ZDV therapy ex vivo. Simple linear regression analysis was used (see Fig. 2 and 3).

FIG. 2.

Relationship between intracellular d4TTP in PBMCs from HIV-infected patients measured in vivo (a) and capacity of PBMCs isolated from HIV⁺ patients to phosphorylate [³H]d4T (1 μM; 5 μCi) ex vivo (b) and time patients had been on a ZDV-containing antiretroviral regimen. Diamonds, ZDV-experienced patients; squares, ZDV-naive patients.

FIG. 3.

Relationship between intracellular d4TTP in PBMCs from HIV-infected patients measured in vivo (a) and capacity of HIV-infected PBMCs to phosphorylate [³H]d4T (1 μM; 5 μCi) ex vivo (b) and length of time since patients were on a ZDV-containing antiretroviral regimen.

RESULTS

Validation of d4T assay.

The recovery of known amounts of d4TTP in the presence of cell extracts varied between 97 and 109%. Consistent inhibitory curves with high reproducibility were obtained with correlation coefficients (r²) greater than 0.98. The lower limit of detection was 0.025 pmol. The coefficient of variation for repeated analysis of d4TTP from the same extract was less than 10%. The observed values are similar to those previously reported using LCMS (Sommadossi et al., 5th Conf. Retrovir. Opportunistic Infect.). Intracellular levels of d4TTP, like the triphosphates of the other nucleoside analogues, show high inter- and intraindividual variation (18, 20; D. J. Back, P. G. Hoggard, S. E. Gibbons, J. Lloyd, M. G. Barry, S. H. Khoo, and C. Loveday, 12th World AIDS Conf., abstr. 42260, 1998; Sommadossi et al., 5th Conf. Retrovir. Opportunistic Infect.).

d4TTP in PBMCs from HIV-infected individuals.

d4TTP concentrations were highest 2 to 4 h after dosing (0.21 ± 0.14 pmol/10⁶ cells; n = 27) (Fig. 1) and were still detectable at 12 h postdose in the one patient analyzed. Comparison of d4TTP concentrations (2 to 4 h postdose) in ZDV-naive and ZDV-experienced individuals showed no significant difference (ZDV naive, 0.23 ± 0.17 pmol/10⁶ cells [n = 7], versus ZDV experienced, 0.20 ± 0.14 pmol/10⁶ cells [n = 20]; P = 0.473). Similarly, there was no significant difference in dTTP levels (ZDV naive, 1.91 ± 0.90 pmol/10⁶ cells [n = 7] versus ZDV experienced, 2.73 ± 1.81 pmol/10⁶ cells [n = 20]; P = 0.761) or the d4TTP/dTTP ratio in ZDV-naive and ZDV-experienced individuals (0.14 ± 0.12 [n = 7] and 0.10 ± 0.08 [n = 20], respectively; P = 0.391).

FIG. 1.

Intracellular d4TTP concentrations (pmol/10⁶ cells) in PBMCs from HIV-infected patients following isolation from a single blood sample at indicated times postdose. The data are expressed as means ± standard deviations.

Incubation of PBMCs with ³H-nucleoside analogues.

Total intracellular phosphate formation in resting and 72-h PHA-stimulated PBMCs is shown in Table 1. Following cell activation, significant increases were seen in the total phosphorylation of d4T (1 μM), ZDV (0.1 μM), and thymidine (1 μM) in PBMCs isolated from both healthy volunteers and HIV patients. However, there was marked intersubject variation. The extent of total phosphorylation of ZDV and d4T was lower in resting PBMCs from healthy volunteers than in those from HIV patients. Formation of d4TTP was also lower in resting PBMCs from healthy volunteers (Table 1). Following PHA activation, the total phosphorylation of thymidine and ZDV was significantly greater in healthy volunteers than in HIV patients (Table 1).

TABLE 1.

Intracellular triphosphates and total phosphates of d4T (1 μM), ZDV (0.1 μM), and ribothymidine (1 μM)^a

PBMCs and phosphate	Intracellular nucleosides (pmol/106 cells)
	Healthy volunteers	HIV patients
Resting PBMCs
d4TTP	0.005 ± 0.002	0.008 ± 0.005*
d4T total phosphates	0.056 ± 0.016	0.083 ± 0.0463*
ZDVTP	0.001 ± 0.001	0.002 ± 0.002
ZDV total phosphates	0.032 ± 0.024	0.081 ± 0.059*
ThdTP	0.021 ± 0.017	0.020 ± 0.021
Thd total phosphates	0.162 ± 0.052	0.232 ± 0.233
PHA-stimulated PBMCs
d4TTP	0.023 ± 0.013	0.020 ± 0.013
d4T total phosphates	0.638 ± 0.313	0.343 ± 0.204*
ZDVTP	0.054 ± 0.057	0.020 ± 0.011
ZDV total phosphates	4.043 ± 2.172	2.796 ± 2.824
ThdTP	0.198 ± 0.157	0.082 ± 0.060
Thd total phosphates	7.553 ± 5.486	2.843 ± 2.691*

^aFollowing 24-h ex vivo incubation in resting and PHA (10 μg/ml; 72 h)-stimulated PBMCs from healthy volunteers and HIV patients. Data are expressed as mean ± standard deviations (n ≥ 6). The data were analyzed by the Mann Whitney U test. ∗, significance between healthy volunteers and HIV patients (P < 0.05). Thd, ribothymidine. 

d4T phosphorylation in resting HIV-infected PBMCs.

d4T phosphorylation in resting PBMCs from HIV patients was further investigated to assess whether prior experience of ZDV reduced anabolite formation. However, no significant difference was seen in total intracellular d4T phosphates in cells isolated from ZDV-naive and ZDV-experienced HIV patients (0.086 ± 0.055 pmol/10⁶ cells in ZDV-naive patients [n = 17] versus 0.081 ± 0.038 pmol/10⁶ cells in ZDV-experienced patients [n = 22]; P = 0.767). Similarly, there was no significant difference between the d4TTP concentrations in these two groups (0.008 ± 0.005 pmol/10⁶ cells in ZDV-naive patients [n = 17] versus 0.008 ± 0.005 pmol/10⁶ cells in ZDV-experienced patients [n = 22]; P = 0.597).

Effect of ZDV experience on d4T phosphorylation in vivo and ex vivo.

The effect on d4T phosphorylation of the length of time that ZDV-experienced patients had been on a combination regimen including ZDV is illustrated for the in vivo (Fig. 2a) and ex vivo (Fig. 2b) studies. In both instances, there was no significant change in d4TTP with either time on ZDV therapy or time since ZDV therapy (in vivo multiple correlation coefficient (r²) = 0.0016, P = 0.997; ex vivo r² = 0.0126, P = 0.875). Figures 2 and 3 illustrate the relationships between d4TTP and these parameters. The ex vivo data also demonstrated that there was no correlation between total d4T phosphates and either time on ZDV or time since ZDV therapy was last administered (multiple correlation coefficient r² = 0.0255; P = 0.703).

Stratification into two groups based on the CD4 count (< 200 versus > 200/μl) or virus load (< 400 versus > 400 copies/ml) demonstrated no significant differences in d4TTP concentration with either time on ZDV or time since ZDV. These data illustrate the fact that differences in CD4 count and virus load did not significantly affect the relationship between d4TTP and duration on, or interval since, discontinuing ZDV.

DISCUSSION

Since there has been considerable concern surrounding the correct sequence for the use of nucleoside analogues, we designed this study to look in the clinical setting for differences in intracellular phosphorylation of d4T in ZDV-naive patients as well as those who had previously received ZDV for various amounts of time. In vivo assessment of PBMCs from patients receiving d4T revealed no differences in formation of d4TTP between ZDV-naive and ZDV-experienced patients. In addition, no difference was observed in the amount of endogenous dTTP or the ratio of intracellular d4TTP to dTTP. Furthermore, there was no relationship between the amount of d4TTP or dTTP formed and either the prior duration of ZDV treatment or the time elapsed since ZDV was discontinued. As patients switched from drug therapy combinations including ZDV to d4T, as shown in Fig. 3, the relationship between the time since ZDV therapy and the d4TTP level demonstrates that the effect of time on d4T also does not affect d4TTP formation.

The ex vivo study examined the capacity of PBMCs to phosphorylate d4T. In agreement with the in vivo study, there was no difference in the capacities of resting and PHA-stimulated cells from ZDV-naive or ZDV-experienced patients to phosphorylate d4T. Similarly, no relationship was observed between d4T phosphorylation and the length of time on ZDV or the interval since ZDV was discontinued.

In the ALTIS trial (14), treatment-naive patients receiving a combination of d4T plus 3TC experienced a 1.66-log-unit drop in virus load at 24 weeks compared to a 0.55-log-unit drop in ZDV-experienced patients. Since neither population had previously received d4T or 3TC and sequencing did not reveal significant differences in virological resistance patterns, it was suggested that prior exposure to ZDV reduced the potency of the new combination of nucleoside analogues. ALTIPHAR (the pharmacological arm of ALTIS [Sommadossi et al., 5th Conf. Retrovir. Opportunistic Infect.]) reported significantly decreased formation of d4T and 3TC triphosphates in a subgroup of patients who were both ZDV experienced and nonresponders (<1-log-unit drop in virus load; n = 6) compared with another subgroup of treatment-naive patients who were responders (>1-log-unit drop in virus load; n = 10). This gave rise to the speculation that prior ZDV treatment may impair subsequent d4T and 3TC phosphorylation. There are alternative explanations for this finding, e.g., the large inter- and intrapatient variability seen in d4T phosphorylation (18, 20; Back et al., 12th World AIDS Conf., Sommadossi et al., 5th Conf. Retrovir. Opportunistic Infect.). Also, concentrating on a group of nonresponders who had prior ZDV therapy may have selected a group of individuals who are poor metabolizers of nucleoside analogues and for whom treatment is thus at increased risk of failing. Preliminary data suggest a correlation between ZDV and 3TC phosphorylation (C. V. Fletcher, S. P. Kawle, D. Weller, T. N. Kakuda, P. L. Anderson, L. Bushman, R. C. Brundage, and R. P. Remmel, Abstr. 7th Conf. Retrovir. Opportunistic Infect., abstr. 94, 2000) and that the amount of ZDV or 3TC triphosphate is related to the rise in CD4 cells and decline in HIV RNA following treatment (C. Fletcher, Program Abstr. First Int. Workshop Clin. Pharmacol HIV Ther., abstr. 6.2, 2000).

In agreement with other studies (8, 9, 15), the formation of ZDV, d4T, and thymidine triphosphates was enhanced with PHA stimulation. This effect was more marked in healthy volunteers than in HIV-positive patients. However, we found that in resting, unstimulated cells, healthy volunteers had lower levels of endogenous nucleotides compared to HIV-positive patients. Several factors may account for these findings. For example, PBMCs from HIV-positive patients may already have high levels of basal activation or may demonstrate a degree of anergy to mitogen stimulation (5). The relative proportions of different cell populations that make up PBMCs also vary with HIV disease, and these subpopulations of cells may respond differently to PHA stimulation. Some of these factors may contribute to the apparently lower levels of thymidine kinase activity seen in PHA-stimulated PBMCs from HIV-infected patients.

Our data highlight the pitfalls of using PHA stimulation when assessing intracellular drug phosphorylation in HIV-positive patients. Moreover, in vitro data from cell lines chronically infected with HIV which suggest reduced phosphorylation with increasing drug exposure (1) are not confirmed. We would caution against extrapolating the findings of such studies to HIV-positive patients and stress the importance of using resting, unstimulated PBMCs in future clinical studies assessing drug phosphorylation.

We have clearly demonstrated that prior use of ZDV does not adversely affect the phosphorylation of d4T, both as a measured amount of drug triphosphate within cells and in regard to the capacity of these cells to phosphorylate added d4T. The choice of whether to use d4T or ZDV when initiating therapy should be based upon the likelihood of toxicity in the patient and the risk of primary drug resistance or subsequently generating cross-resistance to other nucleoside analogues rather than on drug phosphorylation.