Effect of protease inhibitors on nucleoside analogue phosphorylation in vitro

Abstract

Aims

Combination antiretroviral therapy for human immunodeficiency virus (HIV) infection now involves both nucleoside analogues and protease inhibitors. Since intracellular phosphorylation is essential for the activity of all the nucleoside analogues this study was designed to investigate interactions with protease inhibitors at the intracellular level which may alter antiviral efficacy.

Methods

PHA-stimulated PBMCs (3 × 10⁶ cell/plate) and U937 cells (4 × 10⁶ cells/plate) were incubated with either radiolabelled zidovudine (ZDV), stavudine (d4T), zalcitabine (ddC), lamivudine (3TC) or didanosine (ddI) in the presence and absence of the protease inhibitors, indinavir, ritonavir, and saquinavir (0.1−10 μm) for 24 h. Cells were extracted overnight prior to analysis by radiometric h.p.l.c. Intracellular phosphates were standardised to pmol per million cells.

Results

None of the three protease inhibitors tested had any significant effect on the intracellular phosphorylation of the five nucleoside analogues. It is particularly important to focus on the active triphosphate anabolites and data for control vs ritonavir (10 μm) incubations in U937 cells were as follows: ZDVTP, 0.19±0.02 vs 0.21±0.02 pmol/10⁶ cells (mean±s.d.; n = 5); d4TTP, 0.30±0.13 vs 0.27±0.26; 3TCTP, 0.32±0.12 vs 0.26±0.19; ddCTP, 0.07±0.04 vs 0.06±0.02; ddATP, 0.014±0.003 vs 0.018±0.006 pmol/10⁶ cells.

Conclusions

The protease inhibitors, indinavir, ritonavir and saquinavir have no effect on the enzymes responsible for phosphorylation. Combining protease inhibitors and nucleoside analogues should not lead to any intracellular interactions in vivo.

Introduction

Combination antiretroviral therapy is now becoming the standard treatment of human immunodeficiency virus (HIV) infection [1]. At present the two major targets of intervention by antiretroviral agents are the viral reverse transcriptase and protease enzymes.

The dideoxynucleoside analogues are metabolised to their putative triphosphates by intracellular kinases [2]. The nucleoside analogue triphosphates then inhibit the formation of complementary DNA from single-stranded HIV RNA mediated by reverse transcriptase [3]. Protease inhibitors mimic the tetrahedral transition-state intermediate of normal polyprotein substrates. This disrupts the protease enzyme and results in the production of immature non-infectious virions.

The currently approved protease inhibitors undergo metabolism by both intestinal and hepatic CYP3A4 and a number of clinically important drug interactions have been demonstrated [4, 5]. Pharmacokinetic drug interactions involving the nucleoside analogues also occur [2], but since intracellular activation of the nucleoside analogues is essential for their antiretroviral activity it is important to examine the potential for intracellular interactions. Drug interactions at the phosphorylation level have previously been studied with zidovudine (ZDV) [6, 7], stavudine (d4T), [8, 9], zalcitabine (ddC) [10] and lamivudine (3TC) [11]. However to date there have been no studies of the effect of protease inhibitors on the enzymes responsible for phosphorylation of the nucleoside analogues.

The aim of this study was to investigate the potential of the protease inhibitors, indinavir, ritonavir and saquinavir, to interact with the intracellular phosphorylation of the nucleoside analogues; ZDV, d4T, ddC, 3TC and ddI.

Methods

Materials

Lymphoprep was purchased from Nycomed Pharma AS, Oslo, Norway. Foetal calf serum was acquired from Sera, Sussex, U.K. 3TC and [5′-³H]-3TC were kindly donated by Glaxo-Wellcome, U.K. ddI, [2′,3′-³H]-ddI, d4T and [methyl-³H]-d4T were gifts from Bristol-Myers Squibb, Wallingford, CT, U.S.A. ddC, [2′,3′-³H]-ddC and saquinavir were gifts from Roche products Ltd, Welwyn Garden City, U.K. Ritonavir and indinavir were donated by Abbott, IL, U.S.A. and Merck, West Point, PA, U.S.A. respectively. ZDV was donated by MRC AIDS Research Project, Herts., U.K. [methyl-³H]-ZDV was purchased from Moravek Biochemicals, Brea, CA, U.S.A. All other drugs and chemicals were purchased from Sigma Chemical Company Ltd, U.K. [³H]-nucleoside analogues were of ≥97% purity.

Incubation of PBMCs with [3H]-nucleoside analogues

PBMCs (3 × 10⁶cells) were isolated and stimulated with phytohaemagglutinin (PHA; 10 μgml⁻¹) as previously described [6]. Stimulated PBMCs were incubated with either [³H]-ZDV (0.65 μCi, 0.1 μm), [³H]-d4T (0.65 μCi, 3 μm), [³H]-ddC (1.30 μCi, 0.1 μm), [³H]-3TC (0.65 μCi, 1 μm) or [³H]-ddI (3.25 μCi, 0.1 μm). The effects of ritonavir, indinavir and saquinavir on the phosphorylation of the nucleoside analogues were studied by addition of 40 μl of 10 μm, 100 μm and 1 mm solutions to each plate to give final drug concentrations of 0.1 μm, 1 μm and 10 μm respectively. To control plates vehicle was added. Experiments were carried out for 24h at 37° C in a humidified 5% CO₂ gassed incubator and performed in triplicate on five separate occasions.

Incubation of U937 cells with [3H] nucleoside analogues

U937 cells were maintained in 75cm² flasks containing medium supplemented with 10% foetal calf serum and l-glutamine (2 mm) at 37° C in a humidified, 5% CO₂ gassed incubator. For drug interaction studies U937 cells (4 × 10⁶ cells) were incubated with a nucleoside analogue and protease inhibitor as described earlier for stimulated PBMCs.

H.p.l.c. analysis

Following cell collection, washing and extraction as previously described [6], cell extracts were reconstituted in ddH₂0 and separated by h.p.l.c. on an anion exchange column (Partisil 10-SAX; 25cm × 4.6 mm). Samples were eluted with increasing buffer concentrations and phosphate metabolites identified by co-chromatography with authentic standards or occasionally by their elution order previously established [6, 9, 10, 11, 12].

Cell viability and statistical analysis

Cell viability in U937 cells in the presence and absence of interacting drugs was assessed by the method of trypan blue exclusion. Trypan blue (0.5% w/v; 4 μl) was added to an aliquot of cell-containing media (20 μl) and the percentage of non-viable cells determined by counting on a haemocytometer the number of cells unable to exclude the dye.

Both PBMC and cell-line intracellular phosphate data were standardised to pmol per million cells. In U937 experiments results were calculated using cell counts taken at the end of the incubations. However PHA causes clumping preventing haemocytometer cell counting at the end of the incubations, therefore in PBMC incubations results were calculated using cell counts taken at the start of the interaction experiments. Data from interaction studies were analysed by analysis of variance (ANOVA) followed by a modified t-test (Bonferroni).

Results

Previous studies have shown that the optimal conditions for intracellular nucleoside phosphorylation in PBMCs occurs following a 72 h pre-incubation with the mitogen PHA (10 μg ml⁻¹) and all phosphates are detectable following a 24 h incubation [6, 9, 10, 11]. The extent of intracellular metabolism of the nucleoside analogues and h.p.l.c. profiles was comparable to those seen previously [6, 9, 10, 11, 12].

The limit of quantification of individual phosphate anabolites for ZDV, ddC and 3TC was 0.01pmol/10⁶ cells, for d4T, was 0.02 pmol/10⁶cells and for ddI was 0.002 pmol/10⁶cells corresponding to h.p.l.c. peak heights three times base line. The coefficient of variation for repeated analysis of phosphorylated nucleoside analogues, extracted from the same sample was less than 10%.

The protease inhibitors, indinavir, ritonavir and saquinavir, (0.1–10 μm) were shown to have no significant effect on the total phosphorylation of the nucleoside analogues ZDV, d4T, ddC, 3TC and ddI in both PHA-stimulated PBMCs and U937 cells (Figure 1). Similarly there was no significant alteration in the formation of individual intracellular phosphates of any nucleoside analogue in both cell populations following incubation with the protease inhibitors. Table 1 shows data with ritonavir for both cell types. Comparative data were obtained with saquinavir and indinavir.

Figure 1.

The effect of the protease inhibitors, ritonavir, (RTV; 10 μm), indinavir (IDV; 10 μm) and saquinavir (SQV; 10 μm) on the intracellular phosphorylation of ZDV, d4T, 3TC, ddC and ddI in PHA-stimulated PBMCs and U937 cells. Each bar represents mean±s.d.; n = 5 for PBMC data, n = 4 for U937 data. Data analysed by ANOVA.

Table 1.

The effect of ritonavir (RTV) on the intracellular phosphorylation of the dideoxynucleoside analogues (ddN), ZDV, d4T, 3TC, ddC and ddI in PHA-stimulated PBMCs and U937 cells. Data expressed as mean±s.d.; n = 5 for PBMC data, n = 4 for U937 data. Data analyzed by ANOVA. MP, DP and TP = monophosphate, diphosphate and triphosphate respectively. There were no statistically significant differences. ddATP is the active anabolite of ddI.

Discussion

Preliminary results from studies of divergent strategies for the treatment of primary HIV infection have shown that three-drug regimens are superior to two-drug combinations or monotherapy. Combination therapy with two nucleoside analogues and a protease inhibitor is being increasingly used in treatment of HIV [1]. To realise the potential benefits of combination therapies a careful analysis of interactions between the drugs is required.

This study was undertaken to investigate whether the protease inhibitors, indinavir, ritonavir and saquinavir, interfere with the enzymes responsible for the activation of the nucleoside analogues. Interactions with phosphorylation are potentially very important as nucleoside analogues are only efficacious in their triphosphate forms.

In the in vitro system used PBMCs were stimulated with the T-cell mitogen PHA which induces a number of cellular enzymes. U937 cells are an immortal cell line expressing high levels of cellular enzymes and do not require stimulation. These cells were used to substantiate the PBMC data.

The intracellular concentration of phosphates formed from different nucleoside analogues should not be directly compared (Table 1), as the drugs were studied at different concentrations and the enzymes responsible for their intracellular activation are different [9]. Cells do not have the same expression of these enzymes and furthermore mitogen stimulation has differential effects on their expression [13]. The two in vitro cell systems used produce levels of nucleoside analogue anabolites greater than seen in vivo but nevertheless allow investigation of effects of other drugs on phosphorylation.

Previous studies on intracellular phosphorylation have shown that some combinations of nucleoside analogues compete for the same enzymes and may be less effective (e.g.: ZDV/d4T and ddC/3TC) [8, 10, 14]. Similarly numerous interactions have been shown to occur with drugs co-administered with the protease inhibitors. Although no effect of protease inhibitors on the phosphorylating enzymes would be expected, some combinations of nucleoside analogues plus protease inhibitors will have a greater and more sustained effect on viral load than others.

The results present a very important negative. The protease inhibitors, indinavir, ritonavir and saquinavir do not have any significant inhibitory effect on intracellular phosphate formation of the five nucleoside analogues. However it should be remembered that for the protease inhibitors to have an effect on the intracellular metabolism of the nucleoside analogues they must be first taken up into the cell. Preliminary studies with [¹⁴C]-saquinavir have clearly demonstrated that this protease inhibitor accumulates inside U937 cells (unpublished observation) allowing the potential for an interaction at the cellular enzyme level.

Although nucleoside analogue phosphorylation is not affected by the protease inhibitors per se, intracellular pharmacokinetics may be altered indirectly by drug interactions prior to cell entry. For example, the major route of ZDV metabolism is hepatic glucuronidation with minor conversion to 3′-amino-3′-deoxythymidine (AMT) by cytochrome P450. Ritonavir induces ZDV glucuronidation and in consequence the area under the curve is reduced by 25% [15]. However interactions between protease inhibitors and ddI (which is catabolised to hypoxanthine and uric acid [2]) or d4T, ddC and 3TC (which are mainly eliminated by the kidney [2]) seem unlikely.

In conclusion, the present study shows that there is no effect of the protease inhibitors on the enzymes responsible for nucleoside analogue phosphorylation.