Inhibition of the Equilibrative Nucleoside Transporter 1 and Activation of A2A Adenosine Receptors by 8-(4-Chlorophenylthio)-modified cAMP Analogs and Their Hydrolytic Products*

Cyclic AMP analogs containing hydrophobic modification of C8 at the adenine ring such as 8-(4-chlorophenylthio)-cAMP (8-pCPT-cAMP) and 8-(4-chlorophenylthio)-2′-O-methyl-cAMP (8-pCPT-2′-O-methyl-cAMP) can penetrate membranes due to their high lipophilicity and directly activate intracellular cAMP effectors. Therefore, these cAMP analogs have been used in numerous studies, assuming that their effects reflect the consequences of direct activation of cAMP effectors. The present study provides evidence that 8-pCPT-modified cAMP analogs and their corresponding putative hydrolysis products (8-(4-chlorophenylthio)-adenosine (8-pCPT-ado) and 8-(4-chlorophenylthio)-2′-O-methyl-adenosine (8-pCPT-2′-O-methyl-ado)) inhibit the equilibrative nucleoside transporter 1 (ENT1). In PC12 cells, in which nucleoside transport strongly depended on ENT1, 8-pCPT-ado, 8-pCPT-2′-O-methyl-ado, and, to a smaller extent, 8-pCPT-2′-O-methyl-cAMP caused an increase of protein kinase A substrate motif phosphorylation and anti-apoptotic effect by an A2A adenosine receptor (A2AR)-dependent mechanism. In contrast, the effects of 8-pCPT-cAMP were mainly A2AR-independent. In HEK 293 showing little endogenous ENT1-dependent nucleoside transport, transfection of ENT1 conferred A2AR-dependent increase in protein kinase A substrate motif phosphorylation. Together, the data of the present study indicate that inhibition of ENT1 and activation of adenosine receptors have to be considered when interpreting the effects of 8-pCPT-substituted cAMP/adenosine analogs.

The second messenger cAMP binds to different types of effectors, i.e. protein kinase A (PKA), 3 the cAMP binding gua-nine nucleotide exchange factors Epac1 and -2, ion channels, and cAMP-regulated phosphodiesterases and thereby exerts its various important effects in cells (for review, see Refs. [1][2][3]. A large number of studies have used membrane-permeable cAMP analogs containing the hydrophobic 4-chlorophenylthio (pCPT) or bromine modification of C 8 at the adenine ring to study the function of intracellular cAMP effectors (4). However, cAMP analogs may affect cells independently from their ability to directly activate cAMP effectors. Recently, 8-pCPT-2Ј-O-methyl-cAMP and its hydrolysis-resistant Sp analog have been shown to inhibit cyclic nucleotide phosphodiesterases (5), which may facilitate activation of PKA and protein kinase G. Another source of cAMP effector-independent effects of cAMP analogs involves the formation of metabolites. Thus, 8-Cl-cAMP inhibits the viability of tumor cell lines (6,7).
Given the important role of cAMP in cell proliferation (1), we initially studied the effect of membrane-permeable Epac-selective and non-selective cAMP analogs on cell proliferation. We found that 8-pCPT-substituted cAMP analogs as well as their potential hydrolysis products strongly inhibit [ 3 H]thymidine incorporation into PC12 and other cells without affecting the cell number. Further experiments demonstrate that this decrease in [ 3 H]thymidine incorporation is due to an inhibition of [ 3 H]thymidine uptake into cells.
The transport of nucleosides into mammalian cells is accomplished by two types of transporters: concentrative nucleoside transporters, which use the Na ϩ gradient to actively transport nucleosides into the cells, and equilibrative nucleoside transporters (ENTs), which transport nucleosides in both directions and toward equal intra-and extracellular adenosine concentration (8,9). ENTs play key roles in nucleoside and nucleobase uptake for salvage pathways of nucleotide synthesis and are also responsible for the cellular uptake of nucleoside analogues used in the treatment of cancers and viral diseases. By regulating the concentration of adenosine available to ubiquitously expressed cell surface adenosine receptors, nucleoside transporters appear to play important regulatory roles in many physiological processes ranging from cardiovascular activity to neurotransmission and immune function (8 -11). Here we show that 8-pCPT-modified cAMP analogs, and more potently, 8-pCPT-modified adenosine analogs, cause inhibition of the equilibrative nucleoside transporter 1 (ENT1) and activation of G s protein-coupled A 2A adenosine receptors (A 2A R) by an ENT1-dependent mechanism.
Fluorescence-activated Cell Sorting Analysis-PC12 cells were serum-starved for 16 h followed by a 48-h incubation in normal growth medium in the presence or absence of 8-pCPTmodified cAMP analogs. Cell cycle distribution was analyzed using a FACSCalibur apparatus (BD Biosciences) as described (13).
Generation of ENT1 Overexpressing HEK 293 Cells-HEK 293 cells were transfected with a plasmid encoding human ENT1 (hENT1) fused to green fluorescent protein (15) using Lipofectamine 2000 according to the instructions of the manufacturer. Stable clones were obtained by G418 selection.
[ 3 H]Thymidine Incorporation into DNA-Cells grown on 96-well plates were incubated with the cAMP/adenosine analogs for the indicated time. Subsequently, 0.1 Ci of [ 3 H]thymidine was added to each well, and the incubation was continued for 4 h. The cells were washed extensively, and the radioactivity incorporated into DNA was determined using a Wallac TriLux 1450 scintillation counter (PerkinElmer Life Sciences) (14).
[ 3 H]Thymidine Uptake-Cells were grown on 24-well plates and were switched to serum-free DMEM 4 h prior to the experiment. After preincubation of the cells with the nucleotide/nucleoside analogs for 30 min, 0.3 Ci of [ 3 H]thymidine was added, and the incubation was continued for 1.5 min (PC12 and renal mesangial cells) or 15 min (HEK 293 cells). The cells were washed with phosphate-buffered saline, and the radioactivity associated with the cells was determined using a Wallac TriLux 1450 scintillation counter. Where indicated, the uptake experiment was performed in Krebs bicarbonate-buffered solution containing 25 mM NaHCO 3 and 116 mM NaCl or 116 mM N-methyl-D-glucamine, 0.8 mM KH 2 PO 4 , 2.5 mM MgCl 2 , 4.6 mM KCl, 1.2 mM MgSO 4 , 2.5 mM CaCl 2 , and 2 mg/ml D-glucose, pH 7.4.
cAMP Assay-Cells were incubated with cAMP/adenosine analogs for the indicated time. The medium was removed, and the cells were extracted with 0.1 M HCl. The suspension was pipetted up and down until the lysate became homogeneous. The lysate was then centrifuged, and the supernatant was transferred for the determination of the concentration of cAMP using a cAMP enzyme immunoassay kit (Cayman Chemical, Ann Arbor, MI) according to the instructions of the manufacturer.
Apoptosis Assay-Cells grown in 24-well dishes were switched to serum-free DMEM, and 8-pCPT-cAMP/ado and/or ADA were added to the medium. After 24 h, apoptotic cells were assessed by Hoechst 33342 staining. Normal nuclei show faint delicate chromatin staining, nuclei at the early stage of apoptosis display increased condensation and brightness, and nuclei at the late stage of apoptosis exhibit chromatin condensation and nuclear fragmentation.
Adenosine Deaminase Assay-The measurement of the deamination of adenosine and 8-pCPT-cAMP/adenosine analogs was performed as described by Martinek (19).
Data Presentation and Statistical Analysis-The data displayed are means Ϯ S.D. All experiments were repeated at least three times with a similar result. Data from different treatment groups were analyzed by the Student's t test for unpaired samples. p values smaller than 0.05 were considered to indicate statistical significance (*) (see Figs. 1-6).

8-pCPT-modified cAMP Analogs Inhibit [ 3 H]Thymidine
Incorporation into the DNA of Different Cell Types-8-pCPT-cAMP, 8-(4-methoxyphenylthio)-2Ј-O-methyl-cAMP, or 8-pCPT-2Ј-O-methyl-cAMP have been reported to inhibit [ 3 H]thymidine or bromodeoxyuridine incorporation into MCF-7, PC12, thyroid, or MCR-5 cells (7,14,20,21). To investigate whether the decrease in [ 3 H]thymidine incorporation is accompanied by a reduction of the cell number, we compared the effects of the cAMP analogs on [ 3 H]thymidine incorporation with those on the viable cell mass using the MTT assay in different cell types.  (Fig. 2, A-C). 8-pCPT-ado tended to exert the strongest inhibitory effect, but this was not statistically significant. Different from the 8-pCPT-modified cAMP analogs, 8-Br-cAMP, another cAMP analog frequently used to activate PKA and Epac in intact cells, as well as 8-Br-ado, had no effect on [ 3 H]thymidine incorporation into PC12 cells (Fig. 2D).

8-pCPT-substituted cAMP/Adenosine Inhibit ENT1
8-pCPT-modified cAMP/Ado Analogs Inhibit ENT1-Because of the discrepancy of the effect of 8-pCPT-modified cAMP analogs on [ 3 H]thymidine incorporation and cell number, we examined whether the inhibitory effect of C 8 -modified cAMP/ado analogs on [ 3 H]thymidine incorporation could be due to inhibition of [ 3 H]thymidine transport into the cells.  Fig. 1, A and B), and N 6 -benzoyl-cAMP (data not shown) did not. In Panc-1 cells, 8-pCPT-cAMP and 8-pCPT-2Ј-O-methyl-cAMP had no effect (supplemental Fig. 1C). Thus, there was a close correlation between the inhibitory effects of 8-pCPT-substituted cAMP/ado analogs on [ 3 H]thymidine uptake and [ 3 H]thymidine incorporation into the DNA.
The data described so far indicated that 8-pCPT-substituted cAMP/ado analogs inhibited the nucleoside transport across

8-pCPT-substituted cAMP/Adenosine Inhibit ENT1
the plasma membrane. To differentiate whether 8-pCPT-substituted cAMP/ado analogs affected equilibrative or concentrative nucleoside transporters, we studied the effect of replacement of Na ϩ in the incubation medium by N-methyl-D-glucamine on [ 3 H]thymidine uptake. As illustrated in Fig. 3C, replacement of Na ϩ in the incubation medium by N-methyl-D-glucamine did not alter [ 3 H]thymidine uptake into PC12 cells, suggesting that the major part of [ 3 H]thymidine uptake in these cells was Na ϩ -independent. The inhibitory effects of 8-pCPTmodifed cAMP/ado analogs on [ 3 H]thymidine uptake were similar in the presence and absence of Na ϩ in the incubation medium (Fig. 3C), indicating that their effects were on equilibrative rather than concentrative nucleoside transporters.
The ENT1-selective inhibitor NBMPR strongly inhibited [ 3 H]thymidine transport into PC12 cells (Fig. 4A), indicating that the nucleoside transport was mainly mediated by ENT1 in these cells. In Panc-1 cells, NBMPR had no effect (Fig. 4B). Thus, there was a close correlation between the effect of NBMPR and 8-pCPT-modified cAMP/ado analogs on [ 3 H]thymidine uptake, suggesting that these cAMP analogs inhibited [ 3 H]thymidine uptake by inhibiting ENT1. Similar to the 8-pCPT-modified cAMP/ado analogs, NBMPR had no effect on the proliferation of PC12 cells as determined by the MTT test despite its inhibitory effect on [ 3 H]thymidine transport as well as [ 3 H]thymidine incorporation.

8-pCPT-modified cAMP/Ado Analogs Elicit Activation of Adenylyl Cyclase and PKA by an ENT1-and A 2A R-dependent
Mechanism in PC12 Cells-Adenosine transport inhibition may increase the amount of endogenous adenosine available to bind cell surface adenosine receptors, leading to adenosine receptor activation (22,23). In PC12 cells, the G s protein-coupled A 2A R is the only functionally relevant adenosine receptor expressed endogenously (24). Its activation elicits an increase of the intracellular cAMP concentration ([cAMP] i ) and activation of PKA. To investigate the possibility that 8-pCPT-modified cAMP/ado analogs may cause activation of A 2A R signaling, we tested their effects on [cAMP] i and PKA substrate motif phosphorylation RRXp(S/T), reflecting PKA activity. 8-pCPT-ado and, to a smaller extent, 8-pCPT-cAMP, caused an increase in [cAMP] i (Fig. 6A). Accordingly, 8-pCPT-ado and 8-pCPT-cAMP elicited increases in PKA substrate motif phosphorylation (Fig. 6, B, C, and E). The increases in [cAMP] i and PKA substrate motif phosphorylation by 8-pCPT-ado were completely inhibited in the presence of ADA (Fig. 6, A and C), which converts adenosine to inosine, a poor activator of adenosine A 2A R (25). Similar results were obtained with the A 2A R antagonist SCH 58261 (Fig. 6, A, B, and E). In contrast, ADA (Fig. 6D) and SCH 58261 did not affect the inhibitory effects of 8-pCPT-ado on [ 3 H]thymidine uptake. 8-pCPT-2Ј-Omethyl-ado, a potential hydrolysis product of 8-pCPT-2Ј-O-methyl-cAMP, also increased PKA substrate motif phosphorylation by an ADA-and SCH 58261-sensitive mechanism, whereas 8-pCPT-2Ј-O-methyl-cAMP had only a small, if any, effect (supplemental Fig. 2). ADA or SCH 58261 had considerable lower inhibitory effects on 8-pCPT-cAMP-induced PKA substrate motif phosphorylation than on the effects of 8-pCPT-ado (Fig.  6E), which can be reconciled to the fact that 8-pCPT-cAMP, different from 8-pCPT-ado, can also activate PKA directly. As expected, PKA substrate motif phosphorylation in response to the A 2A R agonist CGS 21680 was completely inhibited by SCH 58261 (Fig. 6E). NBMPR also caused an increase in PKA substrate motif phosphorylation, which was reduced by SCH 58261 (supplemental Fig. 3), supporting the suggestion

8-pCPT-substituted cAMP/Adenosine Inhibit ENT1
that inhibition of ENT1 leads to A 2A R-dependent activation of PKA. (26). To strengthen the hypothesis that 8-pCPT-cAMP/ado analogs may cause activation of A 2A R, we studied the effects of 8-pCPTmodified cAMP/ado analogs in the presence and absence of ADA on serum withdrawal-induced apoptosis of PC12 cells. As illustrated in Fig. 6F, 8 (Fig. 6F) or SCH 58261 (data not shown). In contrast, the effect of 8-pCPT-cAMP was ADA-and SCH 58261-insensitive. These data are in agreement with the suggestion that 8-pCPT-ado, 8-pCPT-2Ј-O-methyl-ado, and 8-pCPT-2Ј-O-methyl-cAMP caused anti-apoptotic effects due to their ability to activate A 2A R, whereas 8-pCPT-cAMP can elicit this effect by direct stimulation of PKA.

Expression of hENT1 Confers ADA/SCH 58261-sensitve Activation of PKA by 8-pCPT-cAMP/ado Analogs in HEK 293 Cells-
To further investigate the role of ENT1 in activation of adenylyl cyclase/PKA by 8-pCPT-modified cAMP/ado analogs, we studied the effects of 8-pCPT-cAMP and 8-pCPT-ado on PKA substrate motif phosphorylation in parental as well as hENT1overexpressing HEK 293 cells. 8-pCPT-ado caused an increase in PKA substrate motif phosphorylation in hENT1-overexpressing HEK 293 cells, whereas it had only a small, if any, effect in parental cells ( Fig. 7 and supplemental Fig. 4). The increase in PKA substrate motif phosphorylation by 8-pCPT-ado was inhibited by ADA (Fig. 7) or SCH 58261 (supplemental Fig. 4). Thus, the expression of hENT1 conferred ADA/SCH 58261sensitive activation of PKA by 8-pCPT-ado. Measurement of the deaminase activity of ADA revealed that ADA caused deamination of adenosine itself but not of 8-pCPT-cAMP, 8-pCPT-2Ј-O-methyl-cAMP, 8-pCPT-ado, and 8-pCPT-2Ј-Omethyl-ado. This suggests that the inhibitory effects of ADA on 8-pCPT-cAMP/ado-induced increase in PKA substrate motif phosphorylation were not due to deamination of the 8-pCPT-cAMP/ado analogs. These data supported the suggestion that activation of A 2A R signaling by 8-pCPT-cAMP/ado analogs depended on endogenous adenosine.

DISCUSSION
cAMP analogs bearing hydrophobic substituents at C 8 are membrane-permeant, rendering them capable to activate intracellular cAMP effectors in intact cells (4). Therefore, they are widely used for this purpose, assuming that they act mainly by direct activation of cAMP effectors and thus, their biological effects are due to this mechanism of action. In the present study, we found that 8-pCPT-modified cAMP analogs inhibited ENT1 at similar concentrations as required to activate intracellular cAMP effectors. Considering the ubiquitous presence of ENT1, it is feasible to assume that some effects of these cAMP analogs are due to inhibition of ENT1 rather than direct activation of intracellular cAMP effectors.
A frequently used cAMP analog, possessing both 8-pCPT and the 2Ј-O-methyl modification to achieve membrane permeability as well as Epac selectivity, also inhibited the nucleoside transport. In contrast, bromine, chlorine, or OH modification at C 8 -or N 6 -benzoyl modification did not result in cAMP/ ado analogs that inhibited ENT1. Thus, the present study emphasizes the importance of using a battery of cAMP analogs with known potencies for their target proteins to enable dissection between cAMP effector-dependent and -independent effects. In particular, 8-Br-modified cAMP analogs may offer the possibility to activate PKA and/or Epac without simultaneously inhibiting ENT1.
The inhibitory effects of 8-pCPT-cAMP or 8-pCPT-2Ј-Omethyl-cAMP on nucleoside transport were apparently unrelated to their ability to activate cAMP effectors because their potential hydrolysis products, 8-pCPT-ado and 8-pCPT-2Ј-Omethyl-ado, were at least as potent as their corresponding cAMP analogs to inhibit [ 3 H]thymidine transport. It has recently been reported that 8-pCPT-modified cAMP analogs can be hydrolyzed to the corresponding AMP and adenosine analogs by phosphodiesterases and 5Ј-nucleotidases (5,27) and that the hydrolyzable 8-pCPT-substituted cAMP analogs, but not the non-hydrolyzable Sp-8-pCPT-cAMPS analogs, increase cortisol synthesis in adrenal zona fasciculata cells (28) and reduce the proliferation of the protozoan Trypanosoma brucei (27). Thus, the cAMP effector-independent effects of the cAMP analogs in those studies appear to be due to the formation of metabolites. In the present study, however, short term incubation with the hydrolysis-stabilized Sp-8-pCPT-cAMPS or Sp-8-pCPT-2Ј-O-methyl-cAMPS also inhibited [ 3 H]thymidine uptake, suggesting that hydrolysis of 8-pCPT-substituted cAMP analogs was not essential for their inhibitory effect on [ 3 H]thymidine transport. Nevertheless, in tissues with high phosphodiesterases activity, 8-pCPT-modified cAMP analogs may exert some of its effects through the formation of metabolites, in particular during prolonged incubation.
There was a close correlation between the inhibitory effects of 8-pCPT-substituted cAMP/ado analogs on [ 3 H]thymidine incorporation into the DNA of different cell types with their inhibitory effect on [ 3 H]thymidine transport, indicating that the inhibitory effects of 8-pCPT-substituted cAMP/ado analogs on [ 3 H]thymidine incorporation could be due to suppression of nucleoside transport. Notably, the reduced incorporation of [ 3 H]thymidine into the DNA did not reflect changes in cell proliferation. Similar observations were obtained in MCF-7 cells, in which 8-pCPT-cAMP also has a stronger inhibitory effect on [ 3 H]thymidine incorporation into the DNA than on the cell number (7), in K562 cells, in which a panel of protein kinase inhibitors reduces [ 3 H]thymidine uptake and its incorporation into the DNA without affecting cell proliferation (29), and in EOC-20 microglial cells, in which cannabidiol inhibits [ 3 H]thymidine transport and incorporation but not cell proliferation (27). Thus, measuring the incorporation of nucleosides such as [ 3 H]thymidine or bromodeoxyuridine may not reflect DNA synthesis and cell proliferation in the presence of 8-pCPT-cAMP/ado analogs.
p38 mitogen-activated protein (MAP) kinase inhibitors as well as cannabidiol competitively inhibit substrate binding to ENT1 (30,31). Preliminary experiments indicate that 8-pCPT-cAMP and 8-pCPT-ado did not inhibit the binding of [ 3 H]NBMPR to PC12 cells, suggesting that 8-pCPT-substituted cAMP/ado analogs inhibit ENT1 differently from NBMPR or cannabidiol. Thus, 8-pCPT-cAMP/ado-induced inhibition of ENT1 may not involve competitive inhibition of the binding of the substrate. Thus, the mechanism of 8-pCPT-cAMP/ado analog-induced inhibition of ENT1 is presently unclear and could be indirect.
Inhibition of ENT1 may have a number of important biological consequences. The transporters play key roles in nucleoside and nucleobase uptake for salvage pathways of nucleotide synthesis and are also responsible for the cellular uptake of nucleoside analogues used in the treatment of cancers and viral diseases. By regulating the concentration of adenosine available to cell surface adenosine receptors, nucleoside transporters appear to play important regulatory roles in many physiological processes ranging from cardiovascular activity to neurotransmission and immune function (8 -11). Thus, the adenosine uptake inhibitor KF24345 as well as cannabidiol, which have been shown to inhibit ENT1, reduce lipopolysaccharide-induced tumor necrosis factor-␣ response by an A 2A R-dependent mechanism in mice (23,30). This might be due to increasing the amount of endogenous adenosine and A 2A R activation. However, direct evidence that ENT1 inhibition causes activation of A 2A R was missing.
The data of the present study indicate that 8-pCPT-modified cAMP/ado analogs can activate A 2A R signaling in ENT1expressing cells. Thus, 8-pCPT-modified cAMP/ado analogs caused an increase in [cAMP] i , activation of PKA and prevented apoptosis in PC12 cells, in which [ 3 H]thymidine transport was found to highly depend on ENT1. Similar data were obtained for the effect of 8-pCPT-modified cAMP/ado analogs on PKA activity in HEK 293 cells transfected with hENT1, but not in parental cells, in which [ 3 H]thymidine transport was mainly NBMPR-insensitive. The effects of 8-pCPT-ado, 8-pCPT-2Ј-O-methyl-cAMP, and 8-pCPT-2Ј-O-methyl-ado were ablated by SCH 58261, a specific inhibitor of the G s protein-coupled A 2A R, whereas the effects of 8-pCPT-cAMP were mainly SCH 58261-insensitive. The latter can be reconciled to its ability to directly activate PKA independently from A 2A R. Because inhibition of ENT1 leads to an increase of the extracellular adenosine concentration (8,9), a plausible scenario is that 8-pCPT-modified cAMP/ado analogs inhibit ENT1, thereby facilitating activation of A 2A R by endogenous adenosine. In support of this, exogenous ADA mimicked the effects of SCH 58261.
Although it cannot completely ruled out that 8-pCPT-modified cAMP/ado analogs affect A 2A R directly, our data do not support this possibility. First, ADA completely inhibited PKA substrate motif phosphorylation in response to 8-pCPT-ado, although it did not deaminate 8-pCPT-modified adenosine analogs and had no influence on [ 3 H]thymidine uptake. Second, if 8-pCPT-ado would directly activate A 2A R, this effect should not depend on ENT1.
In addition to inhibition of ENT1 and activation of A 2A R, other mechanisms may also contribute to the elevation of [cAMP] i and activation of PKA by 8-pCPT-cAMP/ado analogs. A recent study showing that 8-pCPT-2Ј-O-methyl-cAMP and Sp-8-pCPT-2Ј-O-methyl-cAMPS inhibit certain cyclic nucleotide phosphodiesterases (5), which facilitates activation of protein kinases A and G, provides one mechanism by which 8-pCPT-modified cAMP analogs may elicit biological effects independently from their ability to activate Epac or PKA. Still other yet unknown mechanisms may participate in the ability of 8-pCPT-modified adenosine analogs to stimulate A 2A Rdependent PKA activation.
In summary, the data of the present study indicate that 8-pCPT-substituted cAMP/ado analogs inhibit ENT1 and can elicit activation of A 2A R. Thus, the effects of cAMP analogs bearing 8-pCPT modification might be due to ENT1 inhibition and/or adenosine receptor activation in addition to direct activation on cAMP effectors. In addition, the elucidation of the mechanism of inhibition of ENT1 by 8-pCPT-modified cAMP/