Nucleoside reverse transcriptase inhibitors (NRTIs) are important drugs in the treatment of infection by the HIV-1. The most widely used today are lamivudine (3TC), abacavir (ABC), azidothymidine (AZT), emtricitabine (FTC), and tenofovir disoproxil fumarate (TDF). Today, long-term efficacy, lack of resistance, and appearance and avoidance of toxicity are the main therapeutic challenges. Inadequate suppression of HIV-1 replication remains a major limitation to successful treatment. Moreover, failure of antiretroviral therapy involves a complex interplay of many factors including poor adherence, virological resistance, and pharmacological issues such as protein binding and cellular resistance (Cinatl et al., 1994; Shehu-Xhilaga et al., 2005). As for virological resistance, it is known that reduced entry or increased efflux of anti-HIV drugs could compromise intracellular drug levels, thus favoring the emergence of resistant viruses (Fridland et al., 2000; Turriziani and Antonelli, 2004).

The most important biochemical and pharmacological features that influence intracellular drug concentrations (in addition to oral bioavailability, plasma protein binding, and physiochemical properties of the drug) are the expression and activity of efficient antiviral influx and efflux transporters, specifically for those drugs incapable of freely crossing cell membranes. Given the role played by specific NRTI uptake and efflux transporters (such as P-glycoprotein and multidrug resistance proteins) in the membrane of CD4+ T cells (main targets for HIV-1) and hepatocytes and in the renal epithelial cells involved in the metabolism and excretion of drugs and xenobiotics, it is important to elucidate the possible involvement of these transporter proteins in drug-drug interactions and toxicity mechanisms (Ford et al., 2004; Koepsell, 2004; McRae et al., 2006).

The expression and activity of a wide variety of efflux transporters of NRTIs and other antiretroviral drugs and their association with intracellular drug levels are well described in the literature (Janneh et al., 2007; Köck et al., 2007). It is interesting that specific polymorphisms in efflux transporters and in metabolizing enzymes have been associated with changes in the plasma levels of some anti-HIV drugs (Owen et al., 2006; Rodríguez Nóvoa et al., 2006; Rotger et al., 2007).

As for influx transporters, some members of SLC28 and SLC29 have been associated with NRTI uptake and with anti-cancer nucleoside analogs (Pastor-Anglada et al., 2004; Errasti-Murugarren et al., 2007), although none of the functionally expressed members of SLC28 or SLC29 gene families seem to be involved in the uptake of AZT in immune cells (Purcet et al., 2006; Minuesa et al., 2008). Polyspecific organic anion, cation, and carnitine transporters of the SLC22 family also have been associated with the uptake of some nucleoside and nucleotide analogs and antiviral drugs. To be specific, human organic anion transporters (hOATs) 1 and 3 have been reported to be involved in the uptake of TDF in renal epithelial cells (Cihlar et al., 2001; Uwai et al., 2007). Our previous findings that hOATs are not expressed in immune cells are particularly interesting (Purcet et al., 2006). As for human organic cation transporters (hOCTs), no specific member has been described yet as an NRTI transporter in immune cells, even though it has been suggested an organic cation transport for AZT and 3TC uptake in microglia and the renal brush-border membrane, respectively (Takubo et al., 2000a; Hong et al., 2001). Because we had found previously that hOCTs are well expressed and functionally active in immune cells and highly up-regulated after activation of CD4+ T cells (Minuesa et al., 2008), we were interested in studying the role of hOCTs in the uptake of NRTIs, the possible cross-inhibition between them, and whether these drugs interact and, therefore, inhibit the physiological function of hOCTs.

We present a detailed study of the interaction between NRTIs and hOCT, the uptake properties of these transporter proteins in relation to 3TC, and the drug-drug interaction among 3TC, ABC, and AZT. We used a stably transfected system of Chinese hamster ovary (CHO) cells to elucidate the role of the transporter proteins in pharmacological efficacy, drug-drug interactions, and toxicity mechanisms in excretion tissues.

Materials and Methods

Reagents.N-Methyl-4-phenylpyridinium (MPP+) ([H₃C-³H], 85 Ci/mmol) was purchased from BioTrend (Köln, Germany). ABC ([³H], 0.5 Ci/mmol), 3TC ([5-³H(N)]; 9 Ci/mmol), AZT ([H₃C-³H]; 12.7 Ci/mmol), and metformin (dimethyl-[¹⁴C]; 112 mCi/mmol) were purchased from Hartmann Analytic GmbH (Braunschweig, Germany).

The substrates and inhibitors of organic cation transport tetraethylammonium (TEA), tetrabutylammonium (TBuA), MPP+, ranitidine (Rani), atropine (Atrop), and 1,1′-diethyl-2,2′-cyanine iodide (D-22) were obtained from Sigma-Aldrich (St. Louis, MO). The remaining reagents used for uptake measurements were also purchased from Sigma-Aldrich unless indicated in the text. 3TC, AZT, ABC, FTC, and TDF were obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health (Bethesda, MD).

Stable Transfection of hOCT in CHO Cells and Cell Culture. hOCT1 (GenBank accession number X98322), hOCT2 (GenBank accession number X98333), and hOCT3 cDNAs (GenBank accession number AJ001417, kindly provided by Dr. V. Ganapathy, Augusta, GA) were recloned into the pcDNA5/FRT/TO vector (Invitrogen, Carlsbad, CA). The eukaryotic expression vectors and an empty vector (pcDNA5), used as a control in uptake experiments, were then transfected into the Flp-In-CHO cell line (Invitrogen) using Lipofectamine 2000 transfection reagent (Invitrogen) according to the manufacturer's recommendations and selected for positive clones with 600 μg/ml hygromycin B (PAA Laboratories GmbH, Linz, Austria). The cell lines with the highest transport activity were chosen for further study and routinely cultured in F-12 (Ham's) medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum in the presence of 300 μg/ml hygromycin B. They were maintained at 37°C in a humidified atmosphere containing 5% CO₂.

“Short-Time” Uptake Measurements. To determine the IC₅₀ values of the NRTIs and to assess which of the NRTIs were substrates of hOCTs, [³H]MPP+ and [³H]3TC uptake in CHO-hOCTs and/or CHO-pcDNA5 (empty vector) cells was measured after 1-s (MPP+) and 15-s (3TC) incubations. In brief, after detaching cells with a soft EDTA/HEPES/NaHCO₃ buffer (0.02%/10 mM/28 mM) and resuspending them in a transport solution (1× phosphate-buffered saline with 0.5 mM MgCl₂ and 1 mM CaCl₂, pH 7.4) at 10⁸ cells/ml, 90 μl (9 × 10⁶ cells) was placed at the bottom of four 2-ml tubes (Sarstedt AG & Co., Nümbrecht, Germany) and shaken in a water bath at 37°C. The uptake measurement was then made tube by tube: 10 μl of radioactive solution (containing the appropriate concentration of the corresponding substrate or inhibitors) was placed on the inner wall of each tube approximately 1 cm above the cells. Uptake measurement was started by vortexing the tube, enabling the radioactive solution and the cells to be mixed, and immediately stopped with 1 ml of stop buffer (cold 1× phosphate-buffered saline plus 100 μM quinine solution). The incubation time was determined using a metronome for 1-s measurements or a timer for 15-s measurements. After two centrifugation/washing steps with stop buffer, cells were lysed and solubilized with 200 μl of guanidine thiocyanate (4 M), mixed with 2 ml of scintillation liquid, and put into the scintillation counter to determine the levels of intracellular radioactivity. The advantage of performing this short-time uptake measurement lies in the reduction of the passive diffusion of the radiolabeled substrates used (MPP+, 3TC, ABC, or AZT). For time course measurements, we used the same procedure described above, with incubation times extended up to 3 min. To measure uptake at 0-s incubation, ice-cold stop solution was added to the cells first, and radioactive substrates were added thereafter.

Kinetic and Statistical Analysis. For the IC₅₀ and K_i analyses, data were fitted to the Hill eq. 1, where V is the uptake of [³H]MPP+ (1 s) or [³H]3TC (15 s) in the presence of the inhibitor (a specific NRTI), V₀ is the uptake of [³H]MPP+ (1 s) or [³H]3TC (15 s) in the absence of the inhibitor, I is the inhibitor concentration (nanomolar), and n is the Hill coefficient. The kinetics graphs (uptake velocity of 3TC versus substrate concentration) were fitted using the classic Michaelis-Menten eq. 2, where V₀ and V_max represent initial and maximal transport velocity, respectively (in picomoles per milligram of protein per minute), [S] is the initial substrate concentration (micromolar), and K_m is the substrate concentration at half-maximal transport velocity (micromolar). We also used the equation to calculate the quotient V_max/K_m, which represents the transport efficiency. A paired Student's t test was used for the statistical comparison of experimental data. We also compared which of the fittings for 3TC inhibition curves (two-binding-site competition versus one-binding-site competition) was best. The kinetic and statistical analyses were carried out using GraphPad Prism versions 4.0 and 5.0 software (GraphPad Software Inc., San Diego, CA).

Results

High-Affinity Interaction of NRTIs with hOCTs. Our first objective was to study the interaction of the currently most used NRTIs with the three subtypes of hOCTs we had cloned previously and stably expressed in CHO cells. To do so, we performed 1-s measurements of [³H]MPP+ (12.5 nM) uptake (within the linearity range; Supplemental Fig. S4) in the presence of different low-range concentrations (from 10^-3-10⁴ nM) of ABC, AZT, FTC, and TDF. All tested NRTIs showed a high-affinity interaction with the three subtypes of hOCTs (Fig. 1). For hOCT1, FTC showed the highest affinity (IC₅₀ = 0.020 nM), followed by ABC, AZT, and TDF (Table 1). The mean percentage of MPP+ transport inhibition ranged from 45 to 60%, with a higher value for ABC (>70% inhibition). For hOCT2, all NRTIs also showed a high-affinity interaction; ABC had the highest interaction value (IC₅₀ = 0.041 nM), followed by AZT, TDF, and FTC (Table 1). For hOCT2, the percentage of MPP+ uptake inhibition was no higher than 60% for all drugs. Finally, for hOCT3, all NRTIs again showed high values and a notable percentage of MPP+ uptake inhibition. The drug that interacted with the highest affinity was TDF (IC₅₀ = 0.005 nM), followed by ABC, AZT, and FTC (Table 1).

Interaction between 3TC and hOCT1, 2, and 3: Identification of High- and Low-Affinity Binding Sites. The existence of two different binding sites with hOCTs for the substrates TEA, choline, and MPP+ and three binding sites for the nontransported inhibitor TBuA have been described elsewhere (Gorbunov et al., 2008). Because of the widespread prescription of 3TC and previous evidence that it could be a substrate for hOCTs (Takubo et al., 2000a, 2002), we focused on the interaction between 3TC and hOCTs, by studying the inhibitory effect of a wide range of concentrations of 3TC on radiolabeled MPP+ uptake. The inhibition of MPP+ uptake clearly followed a biphasic curve (Fig. 2). At low concentrations of nonlabeled 3TC, the inhibition was clear for the three hOCT subtypes but did not reach more than 35 to 40% (Fig. 2). It is interesting that the cells were incubated with concentrations of 3TC greater than 10 μM, and inhibition increased to 80 to 85%. In all cases, the K_d values for the high-affinity binding site were in the picomolar range (Table 2), whereas the K_d values for the low-affinity binding site were in the millimolar range (Table 2). A comparison between fitting the data by one-binding site competition versus two-binding site competition indicated that the latter was preferred in all cases, with statistical significance (p < 0.0001).

To determine whether high- and low-affinity inhibition by 3TC is due to an interaction with previously described high- and low-affinity MPP+ binding sites in OCT1 (Gorbunov et al., 2008), we tested the interaction of 3TC using three different MPP+ concentrations (12.5 nM, 125 nM, and 5 μM) (Fig. 3). Results showed that IC₅₀ values for the high-affinity binding site increased at higher concentrations of substrate (MPP+), whereas IC₅₀ values for the low-affinity binding site remained unchanged (Table 3).

hOCTs Show a Saturable Time Course and Facilitate Transport of 3TC. We then tested the uptake of radiolabeled 3TC in CHO cell lines stably expressing hOCTs in comparison with the uptake in CHO cells stably expressing the empty vector (pcDNA5). We found a substantial difference between uptake rates; although CHO-pcDNA5 cells transported very low quantities of 3TC showing a nonsaturable and linear uptake, CHO-hOCT1, -hOCT2, and -hOCT3 cell lines showed a saturable uptake (up to 60–90 s) and high transport of 3TC (Fig. 4). [³H]3TC uptake was shown to be linear for hOCT1 during the first 15 s (inset, Fig. 4a).

The same time course experiments performed with radiolabeled ABC and AZT showed some differences in CHO-hOCT1 versus CHO-pcDNA5 for AZT and in CHO-hOCT1 and -hOCT3 for ABC. Nevertheless, when performing [³H]ABC or [³H]AZT uptake inhibition with hOCTs substrates, we concluded that these proteins were not involved in transport (Supplemental Figs. S1 and S2).

3TC Uptake Can Be Inhibited Either by hOCT Substrates and Their Inhibitors or by Nonradiolabeled 3TC. After demonstrating that 3TC was taken up by the three hOCTs, we assessed whether this uptake could be inhibited by substrates and inhibitors of hOCTs and by the nonradiolabeled drug. Therefore, we performed radiolabeled 3TC uptake in CHO-hOCT cell lines inhibited by the substrate MPP+ (2 mM), by the low-affinity inhibitors TBuA (2 mM), Rani (2 mM), and Atrop (2 mM), by the high-affinity inhibitor D-22 (200 μM), and by nonradiolabeled 3TC (2 mM). The inhibition reached for hOCT1 and hOCT2 uptake was greater than for hOCT3 uptake (Fig. 5), although they were all statistically significant (p < 0.005). In the case of hOCT1, the highest inhibition was found in the presence of Rani, Atrop, and D-22. For hOCT2, the highest inhibition was found with TBuA followed by Rani, although the remaining drugs also inhibited at similar levels. hOCT3 showed the lowest inhibition, 60% with D-22 and Rani, and nearly 50% with the remaining compounds. Nevertheless, inhibition was statistically significant in all cases and showed that 3TC was a substrate for hOCTs.

These results were confirmed by uptake experiments in X. laevis oocytes injected with cRNA of hOCT1 and hOCT2. Thus, the uptake of [³H]3TC and [³H]MPP+ was inhibited by 2 mM 3TC and 1 mM MPP+ respectively, whereas the uptake of [³H]ABC and [³H]AZT was not affected by 2 mM ABC and AZT, respectively (Supplemental Fig. S3).

Transport Kinetics Revealed hOCT1 to Be the Most Efficient 3TC Transporter of hOCTs. To kinetically characterize the uptake of 3TC by the three hOCT subtypes and compare the efficacy of transport and affinity for the substrate, we performed 3TC uptake at increasing concentrations of the drug (Fig. 6). All three hOCTs showed saturable transport kinetics and followed a Michaelis-Menten curve for 3TC uptake, with K_m in the same order of magnitude (millimolar range) and only slight differences in V_max values (Table 4). To be specific, hOCT1 showed the highest affinity for the 3TC substrate, with a saturation curve with an estimated K_m of 1.25 ± 0.1 mM and a V_max of 10.40 ± 0.32 nmol/mg protein/min, followed by hOCT2 and hOCT3. Moreover, hOCT1 also showed the highest efficiency of transport, with a V_max/K_m quotient of 8.33 ± 0.40 μl/mg protein/min, that is, 2-fold higher than hOCT2 and hOCT3, which showed the same transport efficacy (4.10 ± 0.30 versus 4.30 ± 0.30 μl/mg protein/min, respectively).

Finally, we aimed to compare the transport efficiency of 3TC with MPP+ (model substrate of hOCTs) and metformin (an antidiabetic drug recently discovered as an hOCT1 and hOCT2 substrate), a drug with a K_m also in the millimolar range, in our cell system. To do so, we performed kinetic curves with both drugs and we found a very high V_max/K_m for MPP+ (293.5 μl/mg protein/min), as expected, but a 5-fold times lower transport efficiency for metformin (1.7 μl/mg protein/min) than the one found for 3TC (Supplemental Fig. S4).

ABC and AZT Inhibit 3TC Uptake at Very Low Concentrations. Finally, because we had seen that all the NRTIs interacted with a high affinity with hOCTs (Fig. 1) and that 3TC was a substrate for these transporters, we wanted to explore whether other NRTIs usually taken in combination with 3TC during highly active antiretroviral therapy (HAART) could inhibit 3TC uptake in our stably transfected cell system (Fig. 7). Using low concentrations of ABC and AZT (up to 10 μM), we found inhibition of hOCT-mediated transport of 3TC, with K_i ranging from 6.2 ± 4.1 pM (ABC inhibition for hOCT2-mediated transport) to 330 ± 280 pM (ABC inhibition for hOCT1-mediated transport) (Table 5). It is important that both NRTIs showed considerably different affinities for the transporter subtypes; as an example, ABC had a more than 50-fold greater affinity for hOCT2 and 5-fold greater affinity for hOCT3 than for hOCT1. Nevertheless, for both drugs, no inhibition was more than 45 to 55% at the highest concentration tested.

Discussion

The present study assesses the interaction of NRTIs with hOCTs and cell uptake transport of three of the most widely prescribed NRTIs: 3TC, ABC, and AZT. The study shows that 1) 3TC is a substrate for hOCTs, with hOCT1 being the most efficient transporter; 2) all tested NRTIs interact with hOCTs with a high affinity, and transporter activity is inhibited by 50 to 70%; and 3) ABC and AZT (NRTIs frequently coadministered with 3TC) inhibit the transport of 3TC by up to 70% through hOCTs at low concentrations.

Even though some drug uptake transporters have been described in renal epithelial cells and hepatocytes, it is still unknown which ones are implicated in NRTI transport across biplasma membranes of immune cells. The SLC22 gene family, which encodes for hOAT and hOCT proteins, has been associated with the uptake of antiviral drugs used in HIV and other viral infections (Chen and Nelson, 2000; Takeda et al., 2002). We previously determined that hOCTs are highly expressed, active, and up-regulated in immune cells (including CD4+ T cells, main targets of HIV) (Minuesa et al., 2008). Here, we first studied the interaction of hOCTs with NRTIs, finding a high affinity (IC₅₀ in the picomolar range) interaction in all cases. This means that all NRTIs (with C_max in plasma ranging from 2–10 μM) could physiologically inhibit the in vivo function of the three hOCT subtypes tested and potentially inhibit the transport of xenobiotics or endogenous substrates (Koepsell et al., 2007). The expression and activity of hOCTs are increasingly important in the clinical response to drugs such as metformin (used to treat type 2 diabetes), imatinib (used against chronic myeloid leukemia), or cisplatin and oxaliplatin (used in chemotherapy against some solid tumors) (Yonezawa et al., 2006; Zhang et al., 2006; Shu et al., 2007; Wang et al., 2008). As a consequence, this high-affinity interaction of NRTIs with hOCTs could play a role in drug-drug interactions.

To our knowledge, this high-affinity interaction of NRTIs with hOCTs is a new finding. Other authors have described the interaction of protease inhibitors (PIs) and the anti-infective drugs pentamidine and trimethoprim with hOCTs, but the IC₅₀ values were much higher (Zhang et al., 2000; Jung et al., 2008). The inhibition effect of both NRTIs and PIs at different concentrations could have implications for clinical outcome in combinations of these drug families during HAART. Nevertheless, further studies are necessary before we can draw conclusions on the possible antagonistic effects of specific drug combinations.

3TC is included frequently in HAART regimens and is one of the first choices in the treatment of therapy-naive patients. Previous studies had suggested that it could be a substrate for OCTs (Takubo et al., 2000a). Therefore, we tested the interaction of the three hOCT subtypes with a wide range of concentrations of 3TC. It is interesting that we found a two-interaction site inhibition curve with a first, high-affinity binding site in the picomolar range and a second, low-affinity binding site in the millimolar range. The presence of two-binding site interactions has recently been shown for choline, TEA, and MPP+ for rat OCT1 (rOCT1) in epifluorescence measurements performed in X. laevis oocytes (Gorbunov et al., 2008). It is noteworthy that high-affinity binding site for 3TC in hOCT1 showed a similar K_d value as the one for MPP+ in rOCT1. The existence of a high-affinity MPP+ binding site in hOCT1, with a K_d in the picomolar range, also has been demonstrated recently (Moaddel et al., 2005). We showed previously that high-affinity binding of the nontransported inhibitor TBuA inhibited MPP+ uptake by an rOCT1 mutant but not by the wild-type transporter (Gorbunov et al., 2008). This study provides evidence that binding of transported or nontransported compounds to high-affinity sites may lead to the inhibition of human wild-type OCTs because all NRTIs investigated partially inhibited uptake of 3TC and/or MPP+ by the hOCTs. Functional, molecular, and structural characterization of rOCTs has provided evidence that OCTs contain a large binding cleft that switches from an outward-facing to an inward-facing conformation during the transport cycle (Volk et al., 2003; Gorboulev et al., 2005). The high-affinity binding sites for organic cations are probably located outside the innermost cavity of the outward-facing binding cleft (Gorbunov et al., 2008). The functional role of the high-affinity binding sites remains unclear. Assuming a rate constant for association (K_on) lower than 1 × 10⁹ M^-1 · s^-1, the K_d values (picomolar) found for NRTIs, suggest half-time for dissociation in the order of hours (Corzo, 2006). Thus, these sites would be occupied during a long time in the presence of low concentrations of individual ligands and might modulate the substrate selectivity for transport and/or the transport velocity. In addition, in the presence of other ligands (e.g., NRTIs), these would release the previously bound molecules by either direct competition at the same binding site or by allosteric effects on other high-affinity binding sites nearby. In competition experiments, when we increased [³H]MPP+ concentrations, IC₅₀ values for 3TC at the high-affinity site experimented a significant shift, an evidence of direct competition between MPP+ and 3TC at these sites.

The similarity between the low-affinity IC₅₀ value for 3TC inhibition and the K_m value for 3TC uptake suggests that the low-affinity inhibition site of 3TC is identical with the transport site for 3TC. Despite this, this site may not overlap largely with the low-affinity binding site for MPP+ because no competition between MPP+ and 3TC was found at the low-affinity binding site of 3TC. In accordance, recent data from our laboratories suggest that different hOCT-transported compounds may have partially different binding regions in the low-affinity binding site (H. Koepsell, unpublished data).

Few studies have investigated the role of organic anion and cation transporter proteins in the uptake of antiretroviral drugs (Cihlar et al., 2001; Uwai et al., 2007). In our study, the direct uptake of [³H]NRTIs indicated that 3TC is a substrate for hOCTs, whereas ABC and AZT are not (Supplemental Figs. S1 and S2). Although AZT influx involves a 40 to 50% protein-associated mechanism in T lymphocytes (Purcet et al., 2006), hOCTs do not seem to be relevant for uptake. For 3TC, the percentage of inhibition with substrates of hOCT, the kinetic parameters (K_m, V_max), and the transport efficiency (V_max/K_m) allow us to conclude that hOCT1 is the best transporter, although hOCT2 and hOCT3 can also participate in its uptake. A recent study has described 3TC and zalcitabine as substrates for hOCT1 and hOCT2 but shows only one interaction site with an IC₅₀ in the low micromolar range (Jung et al., 2008). In relation to kinetic parameters, both studies showed the same transport efficiency values and agreed on the fact that hOCT1 was the most efficient transporter.

Even though our K_m values for 3TC are in the millimolar range (1.25–2.14 mM), and the transport efficiency of OCTs for MPP+ is much higher than that for 3TC (Supplemental Fig. S4), the transport of 3TC by OCTs could be relevant in vivo. This assertion is supported by the fact that genetic polymorphisms in hOCT1, which transports metformin with a K_m of 2.42 mM, play a role in modulating the clinical response to the drug by influencing plasma disposition and pharmacokinetics (Shu et al., 2007). Moreover, in our system, hOCT1 showed a transport efficiency for metformin 5-fold times lower than the efficiency of hOCT1 for 3TC.

Finally, we wanted to assess the issue of drug-drug interactions because some NRTIs are frequently coadministered. Two coformulations are frequently prescribed as first line antiretroviral regimens for HIV-1 infection: Kivexa (ABC and 3TC) and Truvada (FTC and TDF), both in combination with either a PI or non-NRTI. In our study, we focused on ABC and AZT taken separately at low concentrations. As expected, both NRTIs inhibited 3TC uptake by up to 50%. The implications for clinical practice are important because both hOCT1 and hOCT2 are highly expressed in the kidney, and 3TC is mainly eliminated via the kidneys (Epivir drug information sheet; GlaxoSmithKline, Welwyn Garden City, Hertfordshire, UK). ABC and/or AZT could inhibit hOCTs function as a modulator of 3TC renal clearance and pharmacokinetics. This inhibitory phenomenon of 3TC renal clearance and the maintenance of higher levels of 3TC in plasma have been described previously for trimethoprim, a drug widely used against Pneumocystis jiroveci pneumonia in HIV+ patients (Sweeney et al., 1995; Takubo et al., 2000b). Moreover, subsequent clinical trials have confirmed that this interaction might be clinically relevant (Moore et al., 1996). The effect of trimethoprim on the clearance of 3TC, emtricitabine, and apricitabine, with similar chemical structures to 3TC, have been confirmed in rat kidney (Nakatani-Freshwater et al., 2006). Extended studies to clarify the role of hOCTs in the uptake of other NRTIs with similar structures to 3TC are under way in our laboratories.

In conclusion, this study provides evidence of hOCTs as important determinants of 3TC intracellular and plasma concentrations because all three hOCT subtypes transport 3TC and are expressed in both immune cells and excretion tissues. The finding that 3TC is a substrate of hOCTs and that NRTIs are high-affinity inhibitors of hOCT function provides new insights into drug-drug interactions. Because of the coadministration of ABC and AZT with 3TC in HAART, these observations could have important implications for clinical practice, especially with regard to 3TC clearance and pharmacokinetics.