Address correspondence and reprint requests to Dr. F. E. Parkinson at Department of Pharmacology and Therapeutics, University of Manitoba, 753 McDermot Avenue, Winnipeg, MB, Canada R3E 0T6. E-mail: Fiona_Parkinson@umanitoba.ca
Abstract: Adenosine, through activation of membrane-bound receptors, has been reported to have neuroprotective properties during strokes or seizures. The role of astrocytes in regulating brain interstitial adenosine levels has not been clearly defined. We have determined the nucleoside transporters present in rat C6 glioma cells. RT-PCR analysis, 3H-nucleoside uptake experiments, and [3H]nitrobenzylthioinosine ([3H]NBMPR) binding assays indicated that the primary functional nucleoside transporter in C6 cells was rENT2, an equilibrative nucleoside transporter (ENT) that is relatively insensitive to inhibition by NBMPR. [3H]Formycin B, a poorly metabolized nucleoside analogue, was used to investigate nucleoside release processes, and rENT2 transporters mediated [3H]formycin B release from these cells. Adenosine release was investigated by first loading cells with [3H]adenine to label adenine nucleotide pools. Tritium release was initiated by inhibiting glycolytic and oxidative ATP generation and thus depleting ATP levels. Our results indicate that during ATP-depleting conditions, AMP catabolism progressed via the reactions AMP → IMP → inosine → hypoxanthine, which accounted for >90% of the evoked tritium release. It was surprising that adenosine was not released during ATP-depleting conditions unless AMP deaminase and adenosine deaminase were inhibited. Inosine release was enhanced by inhibition of purine nucleoside phosphorylase; ENT2 transporters mediated the release of adenosine or inosine. However, inhibition of AMP deaminase/adenosine deaminase or purine nucleoside phosphorylase during ATP depletion produced release of adenosine or inosine, respectively, via the rENT2 transporter. This indicates that C6 glioma cells possess primarily rENT2 nucleoside transporters that function in adenosine uptake but that intracellular metabolism prevents the release of adenosine from these cells even during ATP-depleting conditions.
Adenosine is an endogenous nucleoside that is present in brain at concentrations of 40-460 nM (Ballarin et al., 1991) during physiological conditions and increases up to 100-fold during ATP-depleting conditions, such as ischemia (Parkinson et al., 2000). Adenosine interacts with a family of four receptor subtypes, and activation of A1 receptors, in particular, is associated with neuroprotection during ischemia (Rudolphi et al., 1992; Schubert et al., 1994; Von Lubitz, 1999). The use of adenosine per se or adenosine receptor agonists as neuroprotective agents is limited by systemic side effects such as decreased heart rate and blood pressure (Von Lubitz, 1999). Inhibitors of cellular adenosine uptake or adenosine metabolism have been proposed as site- and event-specific agents for increasing adenosine levels in the brain interstitium (Geiger et al., 1997), and such compounds are predicted to have fewer systemic side effects than receptor agonists. However, the roles of specific cell types for adenosine uptake, release, and metabolism in brain are poorly characterized.
Adenosine is formed primarily by dephosphorylation of ATP via intermediate formation of ADP and AMP. Extracellular adenosine has been reported to originate either from cellular release of adenosine per se or from extracellular metabolism of adenine nucleotides (Fredholm, 1997). During ischemia, release of adenosine per se is thought to be the quantitatively more important source of extracellular adenosine (Whittingham, 1990). Cellular release of adenosine occurs via nucleoside transpoters, which are membrane proteins that mediate transmembrane fluxes of purine and pyrimidine nucleosides, including adenosine. These nucleoside transporters are broadly categorized into two classes: concentrative and equilibrative (ENT). Concentrative nucleoside transporters, of which six subtypes have been characterized, are Na+-dependent and couple influx of adenosine or other nucleosides to influx of Na+ (Geiger et al., 1997; Cass et al., 1998). Two ENT subtypes have been characterized and cloned. Both transport purine and pyrimidine nucleosides across plasma membranes in a direction dictated by their concentration gradients. The equilibrative transporters are two unique gene products and are functionally differentiated based on their sensitivity to nitrobenzylthioinosine [nitrobenzylmercaptopurine riboside (NBMPR)]: ENT1 (cloned es transporter) is sensitive to low nanomolar concentrations of NBMPR, whereas ENT2 (ei transporter) is relatively insensitive to NBMPR, with IC50 values of >1 μM (Griffith and Jarvis, 1996). Both ENT1 and ENT2 have been reported to have broad distribution in rat brain, including both neurons and astrocytes (Anderson et al., 1999a, b). Many studies have demonstrated that ENT1 can mediate the cellular release of adenosine (Gu et al., 1995; Sinclair et al., 2000); however, similar investigations of ENT2 are sparse.
The objectives of this study in rat C6 glioma cells were to determine the nucleoside transporters present in these cells, to investigate the efflux of purines, and to characterize purine catabolic pathways during ATP-depleting conditions. Astrocytes have been reported to have a greater capacity for adenosine uptake than neurons (Bender and Hertz, 1986); therefore, astrocytes may be an important regulator of adenosine levels during ischemia-like conditions. Rat C6 glioma cells were chosen for these experiments because they exhibit low levels of adenine nucleotide release (Cotrina et al., 1998), a potential nucleoside transporter-independent source of extracellular adenosine. Our results demonstrate (a) that C6 glioma cells express ENT2, which is capable of nucleoside uptake and release, and (b) during ATP depletion, ATP metabolism occurs via an IMP pathway resulting in formation and release of hypoxanthine, rather than adenosine.
MATERIALS AND METHODS
PCR primers, low- and high-glucose Dulbecco's modified Eagle's medium, fetal bovine serum, Moloney murine leukemia virus reverse transcriptase, oligo(dT)12-18, and random-primer DNA labeling kits were purchased from Life Technologies (Burlington, Ontario, Canada). The SNAP RNA isolation kit was purchased from Invitrogen (Carlsbad, CA, U.S.A.). Ready To Go PCR beads were purchased from Amersham Pharmacia Biotech (Piscataway, NJ, U.S.A.). 2-Amino-1,5-dihydro-7-(3-pyridinylmethyl)-4H-pyrrolo[3,2-d] pyridin-4-one (BCX-34; peldesine) was a generous gift from Dr. Philip Morris of Biocryst Pharmaceuticals (Birmingham, AL, U.S.A.). [3H]Adenosine, [3H]formycin B, [3H]adenine, and [3H]NBMPR were purchased from NEN Life Sciences (Mississauga, Ontario, Canada). Silica Gel GF TLC plates were purchased from Fisher Scientific (Whitby, Ontario, Canada). erythro-9-(2-Hydroxy-3-nonyl)adenine hydrochloride (EHNA), dipyridamole (DPR), NBMPR, and nitrobenzylthioguanosine were purchased from Research Biochemicals International (Natick, MA, U.S.A.). All other compounds were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.).
Total RNA was isolated from rat C6 glioma cells using the SNAP RNA isolation kit and treated with DNase I. cDNA synthesis was performed at 37°C for 60 min with a total reaction volume of 60 μl consisting of 300 ng of oligo(dT)12-18 primer, 5 μg of total RNA, 3 mM deoxynucleotide triphosphates, 6.7 μM dithiothreitol, 50 mM Tris-HCl (pH 8.3), 75 mM KCl, 3 mM MgCl2, and 3.3 units of Moloney murine leukemia virus reverse transcriptase. Control reactions were performed by omitting reverse transcriptase.
For PCR, control and reverse transcriptase-treated solutions (2 μl) were amplified using Ready To Go PCR beads. The amplification consisted of 30 cycles of 30 s at 94°C, 30 s at 56°C, and 1 min at 72°C. A final 10-min 72°C elongation step followed, and samples were held at -9°C and then analyzed by electrophoresis on a 1.0% agarose gel. DNA bands were viewed and photographed under UV light following ethidium bromide staining.
rENT1 was amplified with the 5′ primer 5′-CACCATGACAACCAGTCACCAG-3′ and the 3′ primer 5′-TGAAGGCACCTGGTTTCTGTC-3′ to produce a 1.76-kb product. rENT2 was amplified using the 5′ primer 5′-TTACCCAACCTGCACCCTCTC-3′ and the 3′ primer 5′-GTAGCCACATTGCATATGGTGA-3′ to produce a 1.67-kb product (Yao et al., 1997). The presence of mRNA for glyceraldehyde-3′-phosphate dehydrogenase, a ubiquitous housekeeping gene, was used as a loading control and was detected using the 5′ primer 5′-GCTGGGGCTCACCTGAAGGG-3′ and the 3′ primer 5′-GGATGACCTTGCCCACAGCC-3′ to amplify a 343-bp DNA product (bases 346-688) from the rat glyceraldehyde-3′-phosphate dehydrogenase cDNA.
Nucleoside uptake assays
Rat C6 glioma cells were cultured as previously described (Stanness et al., 1997) in 24-well plates until confluent. Cells were washed twice in physiological buffer (118 mM NaCl, 25 mM HEPES, 4.9 mM KCl, 1.4 mM K2HPO4, 1.2 mM MgCl2, 1 mM CaCl2, and 11 mM glucose, brought to pH 7.4 with NaOH) or buffer in which NaCl was replaced with N-methyl-glucamine (NMG). Cells were incubated with [3H]adenosine (1 μM) or [3H]formycin B (10 μM) in 250 μl of Na+ or NMG+ buffer for times ranging from 0 to 300 s. To examine the effect of nucleoside transport inhibitors, cells were exposed to graded concentrations of NBMPR or DPR (1 nM-30 μM) before and during the uptake assays. To examine the effect of ATP depletion, cells were treated with 5 mM iodoacetate (IAA) and 1 mM sodium cyanide (NaCN) in glucose-free buffer for 10 min before [3H]adenosine uptake. For kinetic analysis of [3H]adenosine uptake, cells were incubated with graded concentrations of [3H]adenosine (300 nM-30 μM) for 1 min. To terminate uptake, the extracellular solutions were aspirated, and the cells were rapidly washed twice with ice-cold Na+ or NMG+ buffer. Cellular protein was dissolved by incubating cells overnight with 1 M NaOH (500 μl) at 37°C. Separate aliquots of the dissolved cells were used for protein content determination, using the assay of Bradford (1976), and for liquid scintillation spectroscopy. Uptake values were determined from the radioactivity in the dissolved cells and are expressed as picomoles per milligram of cellular protein using the specific activity of the uptake buffer.
Nucleoside release assays
Release assays were performed with confluent C6 cells cultured in 24-well plates. Cells were washed twice in Na+ buffer and then incubated with [3H]adenine (10 μM) or [3H]formycin B (10 μM) at 22°C. This inwardly directed concentration gradient resulted in tritium accumulation within the cells. After 30 min, cells were washed twice with buffer to remove extracellular tritium. Cellular release of [3H]formycin B was initiated by incubating the cells in fresh buffer, thus providing an outwardly directed concentration gradient. As [3H]adenine is rapidly metabolized to [3H]adenine nucleotides by these cells, cells were treated with glucose-free buffer alone or with the glycolytic inhibitor IAA (5 mM) and the oxidative phosphorylation inhibitor NaCN (1 mM) to stimulate metabolism of [3H]adenine nucleotides and release of the 3H-purine metabolites. Trypan blue exclusion staining indicated no significant difference in viability (≥90%) between buffer-treated and IAA/NaCN-treated cells (data not shown). The effects of the nucleoside transport inhibitors DPR (1 nM-30 μM) and NBMPR (1 nM-30 μM) on 3H-purine release were tested. In addition, the contribution of specific enzymes to the metabolism and release of 3H-purines was examined using 1 μM EHNA to inhibit adenosine deaminase, 100 μM EHNA to inhibit both adenosine deaminase and AMP deaminase (Fishbein et al., 1981), and 10 μM BCX-34 to inhibit purine nucleoside phosphorylase (PNP) (Jurkowitz, et al., 1998). In some experiments cells were treated with IAA and NaCN and then washed and returned to Na+ buffer to mimic a return to normoxic conditions. 3H-Purine release into the extracellular medium was quantified by liquid scintillation spectroscopy and TLC. Cells were dissolved in NaOH and analyzed for protein content. Release data are expressed as picomoles per milligram of cellular protein using the specific activity of the loading buffer.
The method of Schrader and Gerlach (1976) was used to identify the 3H-purines released from C6 cells. In brief, n-butanol, ethyl acetate, methanol, and ammonium hydroxide (7:4:3:4 vol/vol) were mixed, placed in a TLC tank, and allowed to equilibrate for 90 min. Extracellular medium (20 μl) obtained from the nucleoside release assays was spotted onto Silica Gel GF plates with 5 μl of unlabeled carrier. Unlabeled carrier consisted of 15 mM each adenosine, inosine, hypoxanthine, adenine, and AMP, 7.5 mM uric acid, and 6.5 mM xanthine. Plates were chromatographed for 3 h, and samples migrated in the order AMP/uric acid, inosine, xanthine, hypoxanthine, adenosine, and adenine. Spots were outlined under ultraviolet light, scraped off, and transferred to scintillation vials, and 500 μl of 0.2 M HCl was added to each tube. After 1 h, scintillation fluid (5 ml) was added, and radioactivity was determined using scintillation spectrometry.
Adenosine deaminase and AMP deaminase assays
Adenosine deaminase and AMP deaminase assays were performed as previously described (Martinek, 1963; Padua et al., 1990). In brief, C6 cells were homogenized in 0.2 M phosphate buffer (pH 7.4), and homogenate protein content was adjusted to 0.5 mg/ml. Fifty microliters of homogenate was combined with 50 μl of 0.2 M phosphate buffer containing either 20 mM AMP (for AMP deaminase) or 1 mM adenosine (for adenosine deaminase) in the presence or absence of 0.1-500 μM EHNA for 60 min at 37°C. The reactions were terminated with 500 μl of phenol color reagent consisting of 1% (wt/vol) phenol and 0.005% (wt/vol) sodium nitroferricyanide in water. Maximal color from the reaction developed in 15 min at 37°C on addition of 500 μl of a solution containing 0.125 M NaOH and 0.042% sodium hypochlorite. The amount of activity for either enzyme corresponded stoichiometrically (1:1) with the production of ammonia. Standard calibration curves for ammonia production were calculated using ammonium sulfate. AMP deaminase or adenosine deaminase activity was determined by measuring absorbance at 650 nm with a Molecular Devices Emax precision microplate reader.
[3H]NBMPR binding assays
Cells were washed with Na+ buffer and then incubated (22°C) with 0.1-5 nM [3H]NBMPR in the absence or presence of 1 μM nitrobenzylthioguanosine. After a 1-h incubation, cells were washed twice with ice-cold Na+ buffer and then dissolved with NaOH. Samples were analyzed for both tritium and protein content.
Each experiment was performed at least three times in duplicate or triplicate, unless otherwise stated. All results are expressed as mean ± SE values, and statistical significance was determined by ANOVA followed by Bonferroni's post hoc test. Statistical analyses were performed using the software package GraphPad PRISM version 2.0.
To identify the nucleoside transporter subtypes present in rat C6 glioma cells, we performed [3H]adenosine and [3H]formycin B uptake experiments. [3H]Adenosine accumulation in C6 cells was linear for intervals up to 5 min (Fig. 1A-C). This could indicate either concentrative transport or equilibrative transport followed by intracellular metabolism of [3H]adenosine to [3H]adenine nucleotides (see Fig. 5, reaction 7). To distinguish between these possibilities, we first used the poorly metabolizable nucleoside transporter-permeant [3H]formycin B. The accumulation of [3H]formycin B was saturable, indicating accumulation via an equilibrative transporter. [3H]Formycin B accumulation in the C6 cells had a t1/2 of 60 ± 9 s and maximal accumulation of 56 ± 3 pmol/mg of protein. Second, the presence of concentrative Na+-dependent transport was investigated by replacing NaCl in the buffer with NMG+. At intervals up to 5 min, [3H]adenosine accumulation was not significantly affected by Na+ substitution (Fig. 1B). Third, depletion of intracellular ATP is another method of abolishing concentrative nucleoside transport (Thampy and Barnes, 1983; Ohkubo et al., 1991); however, treatment of C6 cells with 5 mM IAA and 1 mM NaCN had no significant effect on [3H]adenosine uptake (Fig. 1C). Thus, no evidence for concentrative, Na+-dependent nucleoside transport was found in these cells. Concentration-dependent accumulation of adenosine (0.3-30 μM) in 1 min had a Km value of 12 ± 1 μM and a Vmax value of 157 ± 23 pmol/mg of protein/min (Fig. 1D).
To determine which ENT subtypes are present in C6 cells, we performed RT-PCR using rENT1 or rENT2 sequence-specific primers, assayed inhibition of [3H]adenosine or [3H]formycin B accumulation by DPR or NBMPR, and tested for specific binding of [3H]NBMPR. RT-PCR analysis of C6 cell total RNA indicated that rENT2 was the predominant nucleoside transporter as a strong band was seen at 1.67 kb using rENT2 primers, whereas a very weak band was seen at 1.76 kb using rENT1 primers (Fig. 2A). NBMPR inhibited the accumulation of [3H]adenosine (1 μM; 1 min) with an IC50 value of 1.05 ± 0.25 μM, whereas DPR had an IC50 value of 119 ± 17 nM (Fig. 2B). Similar results were seen for the inhibition of [3H]formycin B accumulation with IC50 values for NBMPR and DPR of 5.5 ± 0.75 μM and 152 ± 42 nM, respectively (Fig. 2C). Specific binding of [3H]NBMPR was not detected in C6 cells (data not shown). Thus, RT-PCR analysis, IC50 values for NBMPR of ≥1 μM, and the lack of specific binding sites for [3H]NBMPR indicate that rENT2 is the predominant nucleoside transporter in C6 cells.
ENT1 transporters have been reported to mediate both accumulation and release of nucleosides (Gu et al., 1996; Sinclair et al., 2000), but similar data are not available for ENT2. Therefore, we examined the cellular release of [3H]formycin B from C6 glioma cells. A hyperbolic release profile was observed (Fig. 3A), and [3H]formycin B release in 1 min was blocked by DPR and NBMPR with IC50 values of 370 ± 12 nM and 3.3 ± 0.4 μM, respectively (Fig. 3B).
To determine if rENT2 transporters can mediate [3H]adenosine release from C6 cells, we used ATP-depleting conditions, which have been used previously to elevate intracellular adenosine levels and to initiate adenosine release (Daval et al., 1980; Reyes et al., 1995). To radiolabel adenine nucleotides without activating adenosine receptors, we incubated C6 cells with 10 μM [3H]adenine for 30 min. Intracellular [3H]adenine nucleotides accounted for ∼95% of the total tritium within the cells (n = 2). Cells were then treated for 10 min with 5 mM IAA and 1 mM NaCN to block glycolytic and oxidative ATP generation. This resulted in net catabolism of [3H]adenine nucleotides (Ogata et al., 1995). Previously, we used this protocol to stimulate [3H]adenosine release from DDT1 MF-2 smooth muscle cells (Sinclair et al., 2000). This protocol increased the release of 3H-purines from C6 cells; however, NBMPR and DPR had no inhibitory effects on this release (Fig. 4A). TLC analysis of the extracellular medium of IAA/NaCN-stimulated cells indicated that this treatment protocol resulted primarily in the cellular release of [3H]hypoxanthine and not [3H]adenosine (Table 1, experiment 2A). HPLC analysis confirmed that the release of endogenous purines matched the release of 3H-purines (data not shown).
Table 1. TLC analysis of purine release from rat C6 glioma cells
Cells were loaded with 10 μM [3H]adenine for 30 min, then rinsed, and incubated with buffer (experiment 1A), 5 mM IAA and 1 mM NaCN (experiment 2A), or IAA, NaCN, and 10 μM DPR (experiment 3A). After 10 min, solutions were removed and analyzed by TLC for purine content. Cells in experimental experiments 1B and 2B were washed and placed into buffer; cells in experiment 3B were washed and incubated with 10 μM DPR. After 10 min, solutions were again removed and analyzed for purine content. TLC analysis was performed as described in the text. AMP/uric acid and xanthine contents were below detectable levels. Data are mean ± SE values (n = 3), in pmol of purine released/mg of cellular protein.
Statistical significance was determined between experimental groups during the same time interval by ANOVA followed by Bonferroni's post hoc test: ap < 0.05, bp < 0.001 versus experiment 1; cp < 0.001 versus experiment 2.
A recent report using primary cultures of rat astrocytes demonstrated that hypoxia/hypoglycemia induced the release of hypoxanthine; however, returning the cells to normoxic/normoglycemic buffer induced cellular release of both hypoxanthine and adenosine (Ciccarelli et al., 1999). We investigated this phenomenon in C6 cells and found a 465% increase in release of 3H-purines from cells exposed first to IAA/NaCN and then to buffer relative to cells exposed to buffer alone (Fig. 4B). Cells that were only exposed to IAA/NaCN exhibited an intermediate level of purine release: 110 ± 25% increase over buffer-treated cells. DPR had no significant effect on purine release induced by IAA/NaCN treatment, but release evoked by removal of IAA/NaCN was inhibited by ∼50% (Fig. 4B). TLC analysis indicated that the increase in tritium release after removal of IAA/NaCN was due to [3H]inosine (Table 1, experiment 2B); thus, in contrast to the previous study (Ciccarelli et al., 1999), we found no evidence for release of [3H]adenosine during or following IAA/NaCN treatment.
We investigated further the lack of [3H]adenosine release from IAA/NaCN-stimulated C6 cells by examining the roles of adenosine deaminase, AMP deaminase, and PNP in the production and release of 3H-purines. The role of adenosine deaminase was investigated with 1 μM EHNA, which inhibited adenosine deaminase in these cells by ∼90% (n = 2). EHNA (1 μM) alone or in combination with 10 μM DPR did not attenuate either the increase in 3H-purine release during IAA/NaCN treatment or following return to Na+ buffer (n = 4, data not shown). This indicates that [3H]hypoxanthine production was not via intermediary formation of [3H]adenosine but that catabolism occurred by the pathway AMP → IMP → inosine → hypoxanthine (shown in dark arrows in Fig. 5). The role of AMP deaminase was investigated using 100 μM EHNA, which inhibited AMP deaminase by ∼90% (n = 2) as well as adenosine deaminase. EHNA (100 μM) did not significantly affect IAA/NaCN-induced 3H-purine release, suggesting that AMP was metabolized by AMP-preferring 5′-nucleotidase to adenosine (Fig. 5, reaction 6), which exited the cells via rENT2 nucleoside transporters. Following removal of IAA/NaCN, 100 μM EHNA inhibited release by 75% (Table 2, experiment 3B), indicating reduced adenosine formation and release during these conditions. Thus, with AMP deaminase inhibited and ATP-depleting conditions removed, AMP kinase (Fig. 5, reaction 8) activity may have resumed in addition to AMP-preferring 5′-nucleotidase activity. EHNA (100 μM) in conjunction with 10 μM DPR was able to inhibit the increase in 3H-purine release induced during IAA/NaCN treatment and following return to Na+ buffer by 78 and 96%, respectively (Table 2, experiments 4A and 4B). The effect of DPR was greater in the presence of IAA NaCN (Table 2; compare experiments 3A and 4A) than following their removal (Table 2; compare experiments 3B and 4B). This provides further evidence that, in the presence of 100 μM EHNA to inhibit AMP deaminase, a greater amount of adenosine was produced in the presence of, rather than following removal of, IAA/NaCN.
Table 2. Effect of inhibition of adenosine deaminase and AMP deaminase or PNP on IAA/NaCN-evoked 3H-purine release
3H-Purines (pmol/mg of cell protein)
% of control
Cells were loaded for 30 min with 10 μM [3H]adenine, rinsed, and incubated with buffer containing 5 mM IAA and 1 mM NaCN (experiment 1A) or in the added presence of 10 μM DPR (experiment 2A), 100 μM EHNA (experiment 3A), DPR and EHNA (experiment 4A), 10 μM BCX-34 (experiment 5A), or DPR and BCX-34 (experiment 6A). After 10 min, solutions were removed and analyzed for 3H-purines using scintillation spectroscopy. Cells were washed and incubated with buffer (experiment 1B), DPR (experiment 2B), EHNA (experiment 3B), EHNA + DPR (experiment 4B), BCX-34 (experiment 5B), or BCX-34 + DPR (experiment 6B). Note the absence of IAA and NaCN in the second incubation interval. After 10 min, solutions were again removed and analyzed for 3H-purines. Data are mean ± SE values (n = 10 for experiments 1 and 2, n = 4 for experiments 3 and 4, and n = 6 for experiments 5 and 6).
Statistical significance was determined between groups during the same time interval by ANOVA followed by Bonferroni's post hoc test: ap < 0.001 versus experiment 1; bp < 0.001 versus experiment 3; cp < 0.001 versus experiment 5.
PNP consumes inorganic phosphate to convert inosine to hypoxanthine and ribose-1-phosphate. The role of PNP in 3H-purine release was investigated with the cell-permeable PNP inhibitor BCX-34 (10 μM) (Jurkowitz et al., 1998; Morris and Montgomery, 1998; Litsky et al., 1999). Similar to 100 μM EHNA, BCX-34 (10 μM) did not significantly affect the IAA/NaCN-mediated release of 3H-purines (Table 2, experiment 5A) but caused a 35% inhibition of the release that occurred following removal of IAA/NaCN (Table 2, experiment 5B). Thus, with PNP inhibited and ATP-depleting conditions removed, inosine release was decreased, possibly owing to product inhibition of IMP-preferring 5′-nucleotidase (Worku and Newby, 1982). In combination with DPR, BCX-34 inhibited the increase in 3H-purine release both during IAA/NaCN treatment and following return to Na+ buffer by 78 and 91%, respectively (Table 2, experiments 6A and 6B). Thus, the experiments during and following ATP-depleting conditions indicate that [3H]inosine and [3H]hypoxanthine were produced primarily by an AMP deaminase-dependent pathway and not an adenosine deaminase-dependent pathway (Fig. 5, reactions 1 and 2).
A main finding of this study was that rat C6 glioma cells express nucleoside transporters primarily of the ENT2 subtype. Although these transporters are bidirectional, as indicated by both uptake and release of the purine nucleoside [3H]formycin B, ATP-depleting conditions did not induce cellular release of adenosine. Our evidence indicates that AMP was metabolized to IMP via AMP deaminase, IMP was dephosphorylated to inosine by IMP-preferring 5′-nucleotidase, and inosine was phosphorolyzed to produce hypoxanthine and ribose-1-phosphate. Hypoxanthine and lesser quantities of inosine were released from the C6 cells.
C6 cells exhibited a linear rate of accumulation of adenosine for intervals up to 5 min. However, as removal of Na+ or depletion of intracellular ATP had no effect on adenosine accumulation, the apparent concentration of adenosine within C6 cells was likely due to intracellular metabolism and preservation of an inwardly directed concentration gradient for adenosine rather than the presence of Na+-dependent nucleoside transporters. Previously, Na+-dependent nucleoside transport has been reported in primary cultures of rat cerebellar and spinal cord astrocytes (Hosli and Hosli, 1988) but not in cortical or hippocampal astrocytes from human, rat, or chick brain (Thampy and Barnes, 1983; Ohkubo et al., 1991; Gu et al., 1996).
Kinetic constants for adenosine transport have been reported for astrocytes from various species and brain regions. Our data indicate a Km value of 12 μM and a Vmax value of 156 pmol/min/mg of protein. Previously, Km values of 3.4-12 μM and Vmax values of 150-360 pmol/min/mg of protein were reported for primary cultures from the human, mouse, and chick (Hertz, 1978; Thampy and Barnes, 1983; Bender and Hertz, 1986; Matz and Hertz, 1990; Gu et al., 1996). Thus, adenosine transport kinetics appear similar among C6 cells and astrocytes from several species.
Rat brain is known to express high levels of rENT2; most studies report finding 40-60% rENT2, <10% Na+-dependent transport, and the remainder rENT1 (Geiger et al., 1988; Lee and Jarvis, 1988; Johnston and Geiger, 1990; Shank and Baldy, 1990). This study demonstrates that C6 cells have mRNA transcript primarily for rENT2. The RT-PCR data were supported by data from functional assays showing that NBMPR inhibited accumulation of adenosine and formycin B with IC50 values of ≥1 μM. Therefore, C6 cells express almost exclusively rENT2, making these cells a useful model for examining the function and regulation of these transporters.
DPR is less useful than NBMPR in distinguishing ENT1 from ENT2 because it exhibits species differences in potency. DPR inhibits hENT1 at nanomolar concentrations, but micromolar concentrations are usually required to inhibit rENT1 (Griffiths et al., 1997a; Yao et al., 1997). Typically, both DPR and NBMPR have micromolar affinity for h- and rENT2 (Griffiths et al., 1997b; Yao et al., 1997; Crawford et al., 1998), although recently DPR was reported to inhibit hENT2 in transfected cells (Ward et al., 2000) and rENT2 in rat synaptosome (Sweeney et al., 1993) with IC50 values between 150 and 360 nM. In the present study, DPR exhibited higher affinity for inhibition of rENT2 in C6 cells than for recombinant rENT2 expressed in Xenopus oocytes (Yao et al., 1997). The reason for the high affinity of DPR for rENT2 in C6 cells is not known but is currently under investigation. It may indicate multiple isoforms of this nucleoside transporter. The values for NBMPR inhibition of rENT2 reported here are similar to those seen in recombinant systems and cell culture models (Griffith and Jarvis, 1996; Yao et al., 1997). DPR and NBMPR inhibited adenosine uptake into primary mouse cortical astrocytes with IC50 values [DPR, 100-390 nM; NBMPR, 1.64-5.33 μM (Bender and Hertz, 1986, 1987)] similar to those reported here for C6 cells. Thus, nucleoside transporters in mouse cortical astrocytes may closely resemble the ENT2 found in rat C6 cells. In contrast, human fetal astrocytes appear to express both hENT1 and hENT2 transporter activity and exhibit biphasic inhibition curves for DPR and NBMPR (Gu et al., 1996).
Directional symmetry of ENTs has been reported in several tissues containing ENT1 (Gu et al., 1996; Sinclair et al., 2000). Before assessing the role of ENT2 transporters in mediating the cellular release of adenosine during ATP-depleting conditions, we tested for directional symmetry of [3H]formycin B transport. Both uptake and release of [3H]formycin B were evident and were inhibited similarly by NBMPR or DPR. These data indicate a potential role of ENT2 in cellular release of adenosine. However, under ATP-depleting conditions with IAA and NaCN the C6 cells did not release adenosine but instead released hypoxanthine and, to a lesser extent, inosine. It is interesting that adenine (1-100 μM) was unable to inhibit the release of hypoxanthine and that hypoxanthine (10-300 μM) was unable to inhibit the uptake of [3H]adenine (10 μM) in C6 cells (data not shown). This suggests the presence of at least two nucleobase transporters in these cells (Fig. 5, NBT 1 and 2). The nucleobase transporter involved in adenine transport appears similar to that described in human erythrocytes and LLC-PK1 cells (Griffith and Jarvis, 1996). Hypoxanthine release may occur via an equilibrative nucleobase transporter similar to that reported in human placenta (Barros, 1994) or possibly by reversal of Na+-dependent nucleobase transport as has been reported for reversal of Na+-dependent nucleoside transport during conditions that perturb the Na+ gradient (Borgland and Parkinson, 1997).
The pathways involved in IAA/NaCN-induced [3H]hypoxanthine release from rat C6 glioma cells are detailed in Fig. 5: AMP → IMP → inosine → hypoxanthine. Our results are in contrast to previous reports of various stimuli inducing adenosine release from primary rat astrocytes (Caciagli et al., 1989; Ciccarelli et al., 1994, 1999; Ballerini et al., 1995). However, these reports do not differentiate between adenosine release per se and adenine nucleotide release followed by extracellular metabolism to adenosine. As primary cultures of astrocytes can release ATP (Guthrie et al., 1999) and as the nucleoside transport inhibitor propentofylline had no effect on adenosine release (Caciagli et al., 1999), it is possible that the primary cultures of astrocytes released ATP, which was metabolized extracellularly to adenosine. In contrast, C6 cells have been reported to be incapable of ATP release (Cotrina et al., 1998).
Treatment of chick glia with deoxyglucose and oligomycin (Meghji et al., 1989) increased release of adenosine, inosine, and, in particular, hypoxanthine. In our experiments, we did not see increased adenosine release from C6 cells treated with IAA and NaCN (Table 1) or deoxyglucose and dinitrophenol (data not shown). The differences between our results and those with chick glia might be due to differences in nucleoside transporter subtypes or metabolic enzymes expressed in these two cell types. Regardless, data from several studies indicate that hypoxanthine is the major constituent of purine release from astrocytes during ATP-depleting conditions (Meghji et al., 1989; Ciccarelli et al., 1999).
Purine metabolism may differ between C6 cells and primary astrocytes. Although the removal of hypoxic stimulus induced adenosine release in astrocytes (Cicarelli et al., 1999), this was not observed with C6 cells following removal of IAA/NaCN. There was, however, a large increase in inosine release, which was effectively inhibited by DPR. As 1 μM EHNA did not decrease inosine release, it is unlikely that the difference in results was due to higher adenosine deaminase activity in C6 cells compared with primary astrocytes. As has been described for T lymphocytes (Barankiewicz et al., 1990), adenosine release from C6 cells may be precluded by high levels of AMP deaminase (Fig. 5, reaction 2) and/or adenosine kinase (Fig. 5, reaction 7) or low levels of AMP 5′-nucleotidase (Fig. 5, reaction 6). Simultaneous inhibition of AMP deaminase and adenosine deaminase, using 100 μM EHNA, induced release that was sensitive to inhibition by DPR, indicating that adenosine was formed and released when these enzymes were inhibited.
Removal of IAA/NaCN might be expected to enhance purine nucleotide salvage and decrease purine release; in fact, the opposite was observed in both this study and a previous report (Ciccarelli et al., 1999). One possible explanation for the increase in purine release may be that purine salvage enzymes can consume nucleotides and thus may simultaneously promote both salvage and release of purines. For example, conversion of ribose-1-phosphate to phosphoribosyl pyrophosphate by ribosephosphate pyrophosphokinase requires ATP as a pyrophosphate donor. The resulting AMP may enter the pathways illustrated in Fig. 5. Alternatively, reperfusion following simulated hypoxia with metabolic inhibitors can lead to free radical production (Quaife et al., 1991; Myers et al., 1995) that may alter purine metabolism (Spragg et al., 1985) and induce release of inosine and hypoxanthine. The mechanism for the enhanced inosine formation following IAA/NaCN removal is unclear. One possibility is that PNP activity may be decreased due to high phosphoribosyl pyrophosphate or inosine or low Pi concentrations within the cells (Ropp and Traut, 1991).
PNP-mediated production of hypoxanthine may be a protective mechanism during ATP-depleting conditions. Inosine and adenosine have been reported to increase the survival of glial cells (Haun et al., 1996; Jurkowitz et al., 1998) and cocultures of neurons and astrocytes (Litsky et al., 1999) during ATP-depleting conditions; this protection was reduced by inhibition of PNP. PNP activity may maintain the adenine nucleotide pool via production of ribose-1-phosphate, which can enter into glycolysis, and hypoxanthine, which can be converted into IMP via hypoxanthine-guanine phosphoribosyl transferase. Also, release of hypoxanthine by astrocytes may facilitate adenine nucleotide salvage by neurons.
In summary, our results demonstrate that C6 glioma cells do not release adenosine during ATP-depleting events. Unlike primary cultures of astrocytes, C6 cells do not release ATP; thus, our data imply that release of ATP and its extracellular metabolism to adenosine are the primary route of adenosine release from astrocytes. C6 cells metabolized ATP without forming adenosine and were able to accumulate adenosine during ATP-depleting conditions. From these findings, we propose that during ATP-depleting conditions astrocytes may salvage extracellular adenosine derived from ATP released from astrocytes or neurons or from adenosine per se released from neurons. Hypoxanthine released from astrocytes during and following an ischemic insult may be important for adenine nucleotide salvage by neurons.