Transport mechanisms of a novel antileukemic and antiviral compound 9-norbornyl-6-chloropurine.

6-Chloropurines substituted at the position 9 with variously modified bicyclic skeletons represent promising antiviral and anticancer agents. This work aimed to investigate the transport mechanisms of 9-[(1R*,2R*,4S*)-bicyclo[2.2.1]hept-2-yl]-6-chloro-9H-purine (9-norbornyl-6-chloropurine, NCP) and their relationship to the metabolism and biological activity of the compound. Transport experiments were conducted in CCRF-CEM cells using radiolabeled compound ([(3)H]NCP). The pattern of the intracellular uptake of [(3)H]NCP in CCRF-CEM cells pointed to a combination of passive and facilitated diffusion as prevailing transport mechanisms. NCP intracellular metabolism was found to enhance its uptake by modifying NCP concentration gradient. The transport kinetics reached steady state under the conditions of MRP and MDR proteins blockade, indicating that NCP is a substrate for these efflux pumps. Their inhibition also increased the cytotoxicity of NCP. Our findings suggest that the novel nucleoside analog NCP has potential to become a new orally available antileukemic agent due to its rapid membrane permeation.

Due to certain conformational resemblance of NCP to natural nucleosides, where bicyclic scaffold mimics sugar moiety, its transport into the cells via nucleoside transporters (NTs) may be expected as it is frequently observed with pharmacologically active nucleoside analogs 6,7 . NTs belong to solute carrier families 28 and 29 (SLC28 and SLC29), which encode concentrative nucleoside transporters (CNT) and equilibrative nucleoside transporter proteins (ENTs), respectively 8 . The high-affinity translocation of natural nucleosides is mediated by CNTs via the transmembrane sodium gradient, whereas facilitative nucleoside influx and efflux is mediated by ENTs 7 . On the other hand, NCP is considerably more lipophilic compared to physiological nucleosides or traditional synthetic nucleoside analogs, which may favour other transport mechanism such as passive diffusion. Fluid-phase diffusion and receptor-mediated endocytosis have also been reported for 9-(2-phosphonomethoxyethyl)adenine (PMEA) in CCRF-CEM cells and HeLa cells, respectively 9,10 .
The aim of this study was to investigate the mode of intracellular uptake of the novel pseudonucleoside analog NCP to extend our understanding of the usefulness and the limitations of the therapeutic use of the compound. Attention was paid to the interactions with ABC-class transporters such as P-gp or MRP as well as to the interplay between the drug uptake and its intracellular metabolism. Experiments were conducted in CCRF-CEM cell line, as these leukemia cells represent one of the potential target tissues.

Materials
NCP and the glutathione conjugate of NCP were synthesized as described previously 11

Cytotoxicity evaluation
The cytotoxicity of the tested compounds was assessed with XTT cell proliferation kit II (Roche Diagnostics GmbH, Mannheim, Germany) according to the manufacturer's instructions. CCRF-CEM cells were seeded in a 96-well plate at a density of 13 500 cells per well. After 24 h, the tested compounds were added to the culture media and incubated for 72 h before the XTT dye was added. The absorbance at 495 nm was recorded after a 1-h incubation with the dye. Where indicated, CCRF-CEM cells were preincubated with the efflux pump inhibitors (10 mM MK571 or 2 mM LY 335979) for 20 min prior to the addition of NCP. The incubation was done at 37 C in a CO 2 incubator using a rotary stirrer. At indicated time intervals, the uptake process was terminated by centrifugation at 5300 Â g for 1 min through an oil layer (a mixture of silicone and mineral oil at final specific density of 1.05 g/ml). The cells were washed by centrifugation (5300 Â g, 1 min) in PBS, solubilized with Soluene Õ tissue solubilizer overnight and radioactivity was counted in a toluene-based scintillation cocktail. The intracellular volume of CCRF-CEM for the calculation of the actual cytoplasmic concentration of NCP was 3.38 ml/10 7 cells 9 . The HPLC analysis of the NCP metabolites was performed as described previously 5 .
The experiment was conducted in a sodium-free buffer (

Data analysis and statistical procedures
Unless otherwise indicated, the data are presented as mean SD from at least three independent experiments. Statistical evaluation, half-maximal saturation (K M ), maximum rate of uptake (V max ), and IC 50 values were performed using GraphPad Prism v. 5.00 (GraphPad Software, La Jolla, CA). The octanol-water partition coefficient values (log P) were obtained using ChemDraw v. 8.0 (Perkin Elmer, Waltham, MA).

NCP accumulates within the cells and reaches steady state under the conditions of MRP inhibition
Initially, the kinetics of NCP accumulation in the cells was studied. When CCRF-CEM cells were incubated with 10 mM NCP for 0 to 60 min, NCP was accumulated intracellularly to a maximum value of about 600 mM at 10 min after which its level decreased sharply ( Figure 2). To determine the role of major efflux proteins in this decrease, selective MRP and P-gp inhibitors were added. In the presence of the inhibitors, the NCP uptake reached steady state ( Figure 2). The effect of MRP inhibition was considerably higher than that of P-gp, indicating that NCP is preferentially (although not exclusively) a substrate for the efflux pumps of MRP class. The efflux of NCP and/or its metabolites is also supported by the analysis of the extracellular medium, which proved the presence of NCP metabolite, NCP-GS, in considerable amount ( Figure 3). The uptake of NCP was also dependent on its extracellular concentration (0.1-1000 mM) and accumulated in cytosol to a level about sixfold higher than in the medium ( Figure 4). The concentration of half-maximal saturation (K M ) and maximum rate of uptake (V max ) were 0.5 ± 0.1 mM and 4.4 ± 0.4 mmol/l/min, respectively.

Intracellular transport of NCP is energy independent
To examine whether NCP might be transported by an energydependent carrier, we studied its dependence on ATP and Na + level. Two ATP biosynthesis inhibitors (oligomycin and antimycin) were used. Both compounds have been previously shown to inhibit ATP biosynthesis in CCRF-CEM cells at indicated concentrations and time with good efficiency 9 . The transport of NCP was found to be independent on both ATP level and the presence of Na + in the medium (Table 1).

NCP transport is facilitated by its rapid intracellular metabolism
Since the uptake was found to be energy independent while paradoxically displaying saturation kinetics, a possible link between NCP metabolism (GSH conjugation) and transport was investigated. Indeed, NCP transport was nearly completely blocked following the inhibition of GST by ethacrynic acid ( Figure 5A). Similarly, when CCRF-CEM cells were preincubated with GSH-depleting agent buthionine sulfoximine (BSO), the uptake was lower than that of untreated cells ( Figure 5B), suggesting that the intracellular metabolism facilitates NCP membrane transport.

NCP does not use NTs to enter the cell
The involvement of NTs in the transport process was studied in the competition experiments using natural nucleosides. The addition of various purine and pyrimidine nucleosides to an eightfold excess had no effect on the NCP uptake (Table 2). Alternatively, the cells were preincubated with an NT inhibitor dipyridamole prior to the addition of 10 or 100 mM NCP. The NCP intracellular concentrations were 93 ± 9% and 112 ± 7%, respectively, compared to the controls treated with NCP only. Similarly, glucose did not prove to act as a competitive inhibitor of NCP transport either (data not shown).

Microtubule disruption blocks NCP uptake
To characterize further the nature of NCP transport processes, the effects of the microfilament-disrupting agents, cytochalasin B (cytB) and D (cytD), and a microtubule-disrupting drug, colchicine, were tested. While the influence of cytD on the uptake was insignificant ( Figure 6A), when cytB and colchicine were used ( Figure 6A and B), the uptake of NCP dropped up to 47% and 11% of control, respectively. Importantly, no signs of inhibitor toxicity were detectable under the conditions of the experiment. Finally, the influence of a dynamin inhibitor, dynasore, on NCP uptake was studied (Figure 7) to find out whether receptormediated endocytosis plays a role in NCP transport. Its effect on the NCP uptake was also insignificant.

Efflux pumps inhibition increases cytotoxicity of NCP
Since we have shown that NCP-GS is a potential substrate for drug efflux pumps, we also investigated whether P-gp and MRPs    inhibitors sensitize CCRF-CEM to the cytotoxic effects of NCP. It turned out that MK571 (MRP inhibitor) did enhance cytotoxicity of the tested compound more efficiently than LY335979 (P-gp inhibitor) ( Table 3). This is in a good correlation with their differential ability to increase the intracellular concentrations of NCP.

Discussion
NCP represents a class of original pseudonucleoside compounds inspired by carbocyclic nucleosides. NCP was demonstrated to inhibit replication of the Coxsackie B3 virus with considerable efficiency (EC 50 ¼ 0.8 ± 0.2 mM) 1 . Studies in our laboratory also suggested an antitumor potential. In this transport study, NCP accumulated in the cells and exhibited saturation kinetics. As it was clearly expelled from the cells when longer incubation times were used, the mechanism of this efflux was investigated. Blocking the multidrug-resistance proteins efficiently prevented the excretion of NCP and increased the cytotoxicity of NCP to CCRF-CEM cells indicating that NCP is likely a substrate for these transporters.
Although the intracellular concentration of NCP was clearly higher compared to this in the medium, the transport was surprisingly both ATP and Na + independent. It should be noted that the radiometric method used to detect intracellular NCP does not discriminate between the parent compound and its metabolites and the presented intracellular uptake data always represent the sum of both. Therefore, we proposed that the intracellular metabolism of NCP (i.e. GSH conjugation) represents a major driving force of its cellular uptake by modifying the concentration gradient across the plasma membrane. Indeed, both GST inhibitions by ethacrynic acid 13 and GSH depletion by BSO 14 resulted in a significant impairment of the uptake rate. The effect was somewhat less pronounced in BSO-treated cells compared to GST inhibition, which can be explained by the fact that the BSOinduced decrease of GSH in the cells was incomplete (more than 40% of residual GSH following an overnight incubation with BSO) 15 . The conjugation reaction is therefore clearly determining the rate of NCP transport. Our results suggested that the NCP transport mechanism is likely a process not utilizing chemical energy, likely diffusion (passive and/or facilitated) or endocytosis. Since the plots of NCP uptake versus NCP extracellular concentration were nonlinear, simple nonfacilitated diffusion and fluidphase endocytosis should be excluded, because neither of them follow saturation kinetics with respect to the extracellular concentration. This assumption is further supported by the fact that the inhibitor of fluid-phase endocytosis cytD 16,17 had no effect on NCP uptake. On the other hand, reaching the apparent steady-state could also be the effect of the intracellular driving force (i.e. GSH or GST) depletion.
Interestingly, NCP transport was also found to be independent of NTs. According to Damaraju et al. 18 , the wild-type CEM cells possess only hENT1 activity. hENT1 is inhibited by NBTI or dipyridamole 7 . However, dipyridamole did not have any effect on the NCP uptake. In addition, not even high levels of natural NTs substrates did not compete with NCP transport. As for the role of passive diffusion, its contribution to the NCP uptake is heavily supported by its lipophilic nature. NCP is considerably more lipophilic (log P ¼ 2.12) compared to its physiological counterpart adenosine (log P ¼ À1.05) and although the concentrationdependent plot of NCP uptake rather favours the protein-mediated mode of transport, we suggest that passive diffusion also takes place.
Compared to the effect of cytD, cytB and colchicine decreased NCP uptake. Several studies have reported that cytB inhibits the facilitated diffusion of hexoses and nucleosides 19,20 . In contrast, an increase in fluid-phase endocytosis in various cell types treated with cytB was also observed 21,22 . Colchicine has been reported to inhibit the receptor-mediated and nucleoside transport in several mammalian cell lines 16,23 . NCP was not found to be transported by nucleoside or glucose transporters, yet its transport was sensitive to cytB. This may be caused by the binding of cytB to the substrate binding site or to membrane sites in close proximity to the transporters 24 . Apparently, GTP hydrolysis-driven clathrinmediated endocytosis 25 is not involved in the mechanism of NCP transport based on its insensitivity to dynamin inhibitor dynasore 26 . Nevertheless, with respect to the effect of the microtubule network-disrupting agents on NCP transport, a contribution of facilitated diffusion is likely.

Conclusions
Overall, our data point to a combination of passive and facilitated diffusion as general transport mechanisms of NCP uptake.  These are supported by the saturability and ATP-independent character and the sensitivity to the inhibition by cytB and colchicine. Intracellular transport is fast and together with the previously reported high permeability in Caco-2 assay suggests good bioavailability of the compound. Once NCP reaches intracellular compartment, it immediately undergoes a GSTcatalyzed conjugation with GSH. The rapid metabolism further facilitates NCP uptake in the direction of concentration gradient. NCP excretion is mediated by efflux proteins of MDR/MRP type, largely via its polar metabolite NCP-GS.  Table 3. The cytotoxicity of 9-norbornyl-6-chloropurine (NCP) in CCRF-CEM cells in the presence of the inhibitors of drug efflux proteins.