Abstract
Selective inhibitors of adenosine production, degradation and transport were used to potentiate in vivo levels of adenosine and to determine the source of both basal and N-methyl-d-aspartate (NMDA)-induced increases in levels of endogenous adenosine in vivo. Male Sprague-Dawley rats receiving unilateral intrastriatal injections of pharmacological agents were sacrificed 15 min postinjection by high-energy focused microwave irradiation (10 kW, 1.25 s). Ipsilateral and contralateral striata were dissected, and adenosine levels were measured by high-performance liquid chromatography. Inhibition of 5′-nucleotidase by α,β-methylene ADP dose-dependently decreased adenosine levels under basal as well as NMDA-stimulated conditions. Inhibition of nucleoside transport by dilazep and adenosine deaminase by 2′-deoxycoformycin each dose-dependently increased basal adenosine levels. 2′-Deoxycoformycin potentiated NMDA-induced increases in adenosine levels. Inhibition of adenosine kinase by 5′-amino-5′-deoxyadenosine increased basal levels of adenosine, but did not significantly affect NMDA-induced increases in adenosine. 2′-Deoxycoformycin combined with 5′-amino-5′-deoxyadenosine produced a greater enhancement of NMDA-induced increases in levels of adenosine than when either drug was administered separately. Endogenous adenosine in vivoapparently originates from release of adenosine as well as from release and extracellular breakdown of a nucleotide under both basal and NMDA-stimulated conditions. Furthermore, inhibitors of adenosine kinase and adenosine deaminase work best to increase levels of endogenous adenosine under basal and NMDA-stimulated conditions, respectively.
Adenosine, a neuromodulator whose actions are mainly inhibitory, depresses basal and evoked neuronal firing, decreases calcium uptake and inhibits release of excitatory neurotransmitters such as glutamate (Wu et al., 1982; Corradetti et al., 1984; Dunwiddie and Diao, 1994). Any or all of these mechanisms may be responsible for observed neuroprotective effects of adenosine receptor agonists that are mediated through specific cell-surface adenosine receptors (Palmer and Stiles, 1995). Large amounts of adenosine are produced during hypoxia, ischemia and seizure activity where it is thought to have neuroprotective functions (von Lubitz et al., 1995). Adenosine levels which are controlled by metabolic and transport processes govern the degree of receptor activation and neuroprotective effects of adenosine have been mimicked with regulators ofendogenous adenosine levels (REAL agents) (see Geiger et al., 1997). Thus, REAL agents such as inhibitors of adenosine metabolism and uptake may promote neuroprotective actions of adenosine and provide feedback inhibition to limit excitotoxicity and neurodegeneration. One benefit of such a therapeutic approach is that levels may be increased specifically when and where adenosine is being produced and/or released.
Adenosine can be produced by dephosphorylation of AMP by both soluble (cytosolic) and ecto (extracellular)-5′-nucleotidase (EC 3.1.3.5) (Nagata et al., 1984); ecto-5′-nucleotidase activity can be inhibited by α,β-MeADP. Metabolism of adenosine is catalyzed by adenosine kinase (EC 2.7.1.20) which results in formation of AMP or by adenosine deaminase (EC 3.5.4.4) to produce inosine. Adenosine kinase has a much higher affinity (Km = 0.2–2.0 μM) for adenosine than does adenosine deaminase (Km = 20–100 μM). Under basal conditions, the main metabolic route for adenosine appears to be phosphorylation by adenosine kinase, whereas at higher levels of adenosine, adenosine kinase activity is subject to substrate inhibition (Yamada et al., 1980) and adenosine deaminase activity may predominate. The relative contribution of these two enzyme systems has been the subject of much discussion, and selective inhibitors for adenosine kinase and adenosine deaminase have been used effectively as tools to determine the relative contributions of these enzymes in regulating adenosine levels.
Once formed, adenosine can be transported across cellular plasma membranes by bidirectional equilibrative nucleoside transporters or by sodium gradient-dependent active transporters (see Cass, 1995), and nucleoside transporter inhibitors may provide useful information about sites of adenosine formation. For example, if adenosine production is intracellular, then inhibition of bidirectional equilibrative transport is predicted to prevent accumulation of adenosine extracellularly. Alternatively, if adenosine is produced extracellularly by breakdown of a released nucleotide, inhibition of equilibrative transport might be expected to prevent uptake/metabolism of adenosine. However, the use of equilibrative transport inhibitors may not always provide such definitive answers because of ambiguities surrounding adenosine release processes (Clark and Dar, 1989; Geiger and Fyda, 1991; Craig and White, 1993; Fredholm et al., 1994), and the presence of inhibitor-insensitive equilibrative and inhibitor-insensitive sodium gradient-dependent transporters (Cass, 1995).
Previously, we showed that intrastriatal injections of the glutamate receptor agonist NMDA increased levels of endogenous adenosine in striatum and that these levels were potentiated by coadministration of adenosine transport and adenosine deaminase inhibitors (Delaney and Geiger, 1995). Here, we extend that work to determine the sources of adenosine under both basal and NMDA receptor-stimulated conditions as well as an effective strategy to potentiate basal and NMDA-stimulated levels of endogenous adenosine.
Materials and Methods
Animals.
Male Sprague-Dawley rats (170–190 g) were obtained from the University of Manitoba Central Animal Care facility. All procedures followed Canadian Council on Animal Care guidelines and were approved by the Animal Care Committee at the University of Manitoba.
Intrastriatal injections.
Intrastriatal injections and high-performance liquid chromatography analyses of adenosine were performed as described in the accompanying paper (Delaney et al., 1998). Drugs were dissolved in 50 mM Tris-HCl, pH 7.4, except 5′-NH2-5′-dADO which was dissolved in 6 N HCl and pH and volumes were adjusted with 8 N NaOH and 50 mM Tris-HCl.
Chemicals.
Fisher Scientific (Pittsburgh, PA) supplied adenosine. NMDA, 5′-NH2-5′-dADO and α,β-MeADP were obtained from Sigma (St. Louis, MO), and chloracetaldehyde was purchased from Fluka (Ronkonkoma, NY). DCF was a generous gift from Parke-Davis (Detroit, MI). Dilazep was kindly provided to us by Drs. Haefely and Eigenmann at F. Hoffman-La Roche (Basel, Switzerland) and by Drs. Kutscher and Buchner at Asta Pharma AG (Frankfurt, Germany). All other chemicals were of analytical grade and were from standard laboratory sources.
Data analysis.
Adenosine levels in injected striata were calculated as a percentage of levels in uninjected contralateral striata and expressed as mean ± S.E.M. for each drug treatment group. Levels in injected striata were compared with those in the uninjected striata with Student’s paired t tests. Differences between treatment groups were analyzed either by Student’s unpaired t tests or by ANOVA followed by Tukey-Kramer’s multiple-comparison post test. Statistical significance was considered to be at the P < .05 level.
Results
Levels of endogenous adenosine in uninjected striata were 98 ± 16 pmol/mg protein whereas those in striata injected with buffer were 104 ± 3 pmol/mg protein (data not shown). Levels in uninjected striatum calculated as a percent of levels in the corresponding injected contralateral striatum were 97 ± 18%. Intrastriatal injection of 25 nmol NMDA increased significantly (P < .001) levels of adenosine when compared with contralateral striata (238 ± 45%, fig. 1) or with buffer-injected striata (data not shown). Injection of 100 nmol NMDA significantly increased adenosine levels to 464 ± 140%.
The 5′-nucleotidase inhibitor α,β-MeADP (100 nmol) did not significantly affect basal levels of adenosine, but at 200 nmol decreased significantly (P < .05) levels to 17 ± 2% (fig. 1). α,β-MeADP at 100 or 200 nmol reduced significantly (P < .05) NMDA-induced increases by about 47% and 74%, respectively (fig. 1).
In the absence of NMDA, the adenosine deaminase inhibitor DCF dose-dependently increased basal levels of adenosine; significant (P < .01) increases to 233 ± 14% were observed with 40 nmol DCF (fig. 1). DCF potentiated 25 nmol NMDA-induced increases to 380 ± 109% (20 nmol) and to 738 ± 122% (40 nmol) and potentiated 100 nmol NMDA-induced increases to 561 ± 47% (20 nmol) and to 875 ± 215% (40 nmol) (fig. 2). In addition to potentiating levels of adenosine in response to NMDA receptor stimulation, DCF extended the duration of increased levels of adenosine (data not shown). In the absence of DCF, NMDA-induced increases in levels of adenosine declined to basal values by 45 min postinjection (Delaney et al., 1998), whereas in the presence of DCF (20 nmol) adenosine levels remained significantly increased (P < .001, Student’st test, n = 4) to 456 ± 48% 45 min postinjection (data not shown).
Intrastriatal injection of the adenosine transport inhibitor dilazep (2.5 nmol) increased significantly (P < .05) basal levels of adenosine to 349 ± 62% (fig. 3). In the presence of 25 nmol or 100 nmol NMDA, dilazep (2.5 nmol) showed a strong trend to increase NMDA-evoked levels of adenosine to 401 ± 92% and 882 ± 232%, respectively (fig. 3).
DCF (20 nmol) nonsignificantly increased basal levels of adenosine and the adenosine kinase inhibitor 5′-NH2-5′dADO (60 nmol) increased significantly (P < .05) basal levels of adenosine to 155 ± 23% (fig. 4). Coadministration of DCF and 5′-NH2-5′-dADO increased significantly (P < .05) levels of adenosine to 273 ± 59% (fig. 4). 5′-NH2-5′-dADO nonsignificantly increased the levels of adenosine induced by NMDA (25 nmol) from 238 ± 45% to 315 ± 69% (fig. 4). Coadministration of DCF and 5′-NH2-5′-dADO with NMDA significantly (P < .05) potentiated NMDA-induced increases in adenosine levels. DCF also significantly (P < .05) potentiated the effect that 5′-NH2-5′-dADO and NMDA had on levels of adenosine.
5′-NH2-5′-dADO reversed the potentiating effect that dilazep had on basal levels of adenosine; levels of 349 ± 62% elicited by dilazep alone significantly decreased to 170 ± 51% in the presence of 5′-NH2-5′-dADO. Coadministration of 5′-NH2-5′-dADO had no significant effect on NMDA-induced levels of adenosine or those elicited by NMDA and dilazep in combination (fig. 5).
Discussion
Adenosine can be formed intracellularly through the actions of cytosolic 5′-nucleotidase and then exit cells viabidirectional nucleoside transporters, or alternatively, adenosine can be formed extracellularly from metabolism of released nucleotides by ecto-5′-nucleotidase. Although intra- and extracellular compartments in brain tissue obtained from rats sacrificed by microwave irradiation cannot be differentiated, our results here strongly suggest that endogenous adenosine in vivo originated to varying degrees under basal and NMDA-stimulated conditions from release of adenosine as well as from release and extracellular breakdown of adenine nucleotide(s). Furthermore, inhibitors of adenosine kinase and adenosine deaminase worked best to increase levels of endogenous adenosine under basal and NMDA-stimulated conditions, respectively.
With this in vivo approach, the levels of basal adenosine levels (pmol/mg protein) have remained within a 2-fold range for 3 years; levels of 98 and 80 were found here and in the companion study, and values of 143 and 181 were reported previously (Delaney and Geiger, 1996; Delaney et al., 1997). Part of this variability may have been the result of two complete turnovers of rat breeding stocks at our breeding facility. However, more relevant perhaps was our finding during the same period that increases in adenosine levels after NMDA (25 nmol) administration intrastriatally remained within a range of 238 ± 45% to 271 ± 35%.
The ecto-5′-nucleotidase inhibitor α,β-MeADP decreased levels of adenosine under basal conditions by about 83% and under NMDA-stimulated conditions by 47 and 74%. These results strongly suggest that a significant proportion of striatal levels of adenosinein vivo originate from adenine nucleotide metabolism. Previously, in vitro studies showed that inhibition of ecto-5′-nucleotidase activity decreased basal adenosine release by 40% and NMDA-induced adenosine release by 68% (Hoehn and White, 1990;Craig and White, 1993). However, firm conclusions regarding sources of adenosine from such studies apparently depend heavily on the neuronal preparations used as well as the method used to stimulate adenosine release (MacDonald and White, 1985; Lloyd et al., 1993;Cunha et al., 1996).
On the basis of our results with α,β-MeADP, we used dilazep, an inhibitor of equilibrative, presumably bidirectional, nucleoside transporters, to test our hypothesis that adenosine was originating from produced/released adenine nucleotides. It was reasoned that dilazep would decrease measured adenosine levels if intracellular adenosine was released through equilibrative nucleoside transporters (White and MacDonald, 1990; Geiger and Fyda, 1991; Fredholm et al., 1994; Sweeney, 1996). Accordingly, the adenosine remaining inside cells would be metabolized rapidly by adenosine kinase and adenosine deaminase. Conversely, if adenosine originated extracellularly, then inhibition of transport might increase levels by preventing adenosine re-uptake. Although we recognize that the in vivo situation may be far more complex than this (see Geiger and Fyda, 1991; Geiger et al., 1997), our findings that dilazep increased basal levels of adenosine suggests that nucleotide breakdown was occurring under basal conditions and that dilazep increased adenosine levels by blocking re-uptake. Although dilazep did exhibit a trend to increase levels of adenosine evoked by NMDA, this was not statistically significant; the limited solubility of dilazep prevented testing of higher doses. Previously, transport inhibition was reported to increase levels of adenosine under basal conditions (Clark and Dar, 1989; Craig and White, 1993; Gidday et al., 1996), as well as under stimulated conditions (Dunwiddie and Diao, 1994; Fredholmet al., 1994; Gidday et al., 1996) in a variety of in vitro and in vivo preparations.
Adenosine kinase, because of its high affinity for adenosine, may be more important in regulating adenosine levels under basal conditions than under conditions where adenosine levels are increased (see Geigeret al., 1997). Consistent with this, we found that inhibition of adenosine kinase with 5′-NH2-5′-dADO increased basal levels of adenosine. Under NMDA-stimulated conditions, however, adenosine kinase inhibition did not significantly affect adenosine levels, and one explanation for this observation is that concentrations of adenosine were raised by NMDA to levels sufficient to cause substrate (adenosine) inhibition of adenosine kinase, thus rendering the enzyme unresponsive to 5′-NH2-5′-dADO. Alternatively, it may be that the 60 nmol dose of 5′-NH2-5′-dADO used was not high enough to observe a pharmacological result; however, much lower doses were highly effective in preventing convulsions (Zhang et al., 1993). Adenosine kinase inhibition increased basal and evoked release of adenosine and potentiated the actions of adenosine in vitro (Lloyd and Fredholm, 1995; Golembioska et al., 1996; White, 1996) and in vivo (Sciotti and Van Wylen, 1993;Pazzagli et al., 1995; Keil and DeLander, 1996). Adenosine kinase inhibition in two different models of cerebral ischemia produced varying results, exhibiting either neuroprotection (Jiang et al., 1997) or no effect (Phillis and Smith-Barbour, 1993). Thus, the effectiveness of adenosine kinase inhibitors may be linked closely with their condition-specific effects on levels of adenosine.
Inhibition of adenosine deaminase by DCF increased basal and, to a slightly greater extent, potentiated NMDA-induced increases in adenosine levels. Variable results have been noted with DCF or another adenosine deaminase inhibitor,erythro-9-(2-hydroxy-3-nonyl)adenosine, on basal (Lloyd and Fredholm, 1995; Golembioska et al., 1996; White, 1996) and evoked adenosine release in vitro (Lloyd and Fredholm, 1995;Golembioska et al., 1996White, 1996), but consistently increased levels in vivo (Phillis et al., 1991;Sciotti and van Wylen, 1993; Golembioska et al., 1995;Pazzagli et al., 1995). DCF had no effect on antinociception (Keil and DeLander, 1996; Golembioska et al., 1995) or on neurotransmission events (Pak et al., 1994), but had anticonvulsant effects (Zhang et al., 1993) and prevented neuronal damage after ischemia (Phillis and O’Regan, 1989; Lin and Phillis, 1992; Gidday et al., 1995). Thus, the effects of adenosine deaminase inhibitors on the levels and actions of adenosine may be especially condition-sensitive.
Previously, we showed that DCF (20 nmol) in combination with dilazep (0.5 nmol) increased levels of adenosine evoked by NMDA (Delaney and Geiger, 1995). In the current study, we showed that the effects of DCF and dilazep were additive. In contrast, the effects of dilazep and 5′-NH2-5′-dADO on either basal or NMDA-evoked levels were not additive. Indeed, 5′-NH2-5′-dADO decreased the effect of dilazep. This indicates, at least in part, a common mechanism. One explanation may be that, like iodotuberocidin (Parkinson and Geiger, 1997), 5′-NH2-5′-dADO is a permeant for and/or an inhibitor of adenosine transport. Thus, in our experimental system, a combination of transport and adenosine kinase inhibition was not effective at increasing adenosine levels. DCF was shown to be a permeant for nucleoside transporters in human erythrocytes, leukemic and lymphoma cells (Chen et al., 1984; Wiley et al., 1991) and yet showed additive effects with dilazep in the brain, which suggests that DCF may be entering cells through transporters insensitive to dilazep. DCF when combined with 5′-NH2-5′-dADO increased NMDA-evoked levels. Additive affects of inhibition of both adenosine kinase and adenosine deaminase have been found with respect to anticonvulsant activity, cerebral blood flow, adenosine release from rat spinal cord slices andin vivo rat caudate nucleus, (Sciotti and Van Wylen, 1993;Zhang et al., 1993; Golembioska et al., 1995). Therefore, in attempting to fashion a rational therapeutic approach to increase levels of endogenous adenosine in a site- and event-specific manner, inhibition of adenosine deaminase and nucleoside transport or adenosine kinase and adenosine deaminase might lead to the greatest increases in adenosine levels and thus therapeutic benefit.
Footnotes
-
Send reprint requests to: Dr. J. D. Geiger, Department of Pharmacology and Therapeutics, University of Manitoba Faculty of Medicine, 753 McDermot Avenue, Winnipeg, Manitoba, R3E 0T6 Canada.
-
↵1 These studies were supported by a grant from the Medical Research Council of Canada (to J.D.G.).
-
↵2 Recipient of a Medical Research Council of Canada Studentship Award. Current address: Department of Pharmacology, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599.
-
↵3 Recipient of a Medical Research Council of Canada Scientist Award.
- Abbreviations:
- NMDA
- N-methyl-d-aspartate
- α
- β-MeADP, α,β-methylene adenosine diphosphate
- DCF
- 2′-deoxycoformycin
- 5′-NH2-5′-dADO
- 5′-amino-5′-deoxyadenosine
- ANOVA
- analysis of variance
- Received August 28, 1997.
- Accepted January 23, 1998.
- The American Society for Pharmacology and Experimental Therapeutics
References
Levels of endogenous adenosine in striata of rats receiving unilateral injections of 100 nmol or 200 nmol α,β-MeADP in the absence or presence of 25 nmol NMDA (n = 3–10). Levels in injected striata were expressed as a percentage of levels in uninjected contralateral striata. Using pairedt test analyses, levels in injected striata were significantly (* P < .05; *** P < .001) different from levels in uninjected striata. Differences between groups were analyzed by ANOVA followed by Tukey-Kramer’s multiple-comparison post test.
Levels of endogenous adenosine in striata of rats (n = 4–10) receiving unilateral injections of buffer (open circles), 20 (closed circles) or 40 nmol (closed squares) DCF in the absence or presence of 25 nmol or 100 nmol NMDA. Differences between groups were analyzed by ANOVA followed by Tukey-Kramer’s multiple-comparison post test. aP < .01vs. buffer; bP < .01vs. 25 nmol NMDA; cP < .05vs. 25 nmol NMDA/20 nmol DCF.
Levels of endogenous adenosine in striata of rats (n = 4–8) receiving unilateral injections of buffer (open circles), 0.5 nmol (closed circles) or 2.5 nmol (closed squares) dilazep in the absence or presence of 25 nmol or 100 nmol NMDA. Differences between groups were analyzed by ANOVA followed by Tukey-Kramer’s multiple-comparison post test. aP < .05 vs. buffer; bP < .05vs. 0.5 nmol dilazep.
Levels of endogenous adenosine in striata of rats (n = 3–9) receiving unilateral injections of either 60 nmol 5′-NH2-5′-dADO, 20 nmol DCF or 60 nmol 5′-NH2-5′-dADO plus 20 nmol DCF in the absence or presence of 25 nmol NMDA. Using paired t test analyses, levels in injected striata were significantly (* P < .05) different from levels in uninjected striata. Differences between groups with either NMDA absent or present were analyzed by ANOVA followed by Tukey-Kramer’s multiple-comparison post test.
Levels of endogenous adenosine in striata of rats (n = 4–9) receiving unilateral injections of either 60 nmol 5′-NH2-5′-dADO, 2.5 nmol dilazep (DLZP) or 60 nmol 5′-NH2-5′-dADO plus 2.5 nmol dilazep in the absence or presence of 25 nmol NMDA. With paired t test analyses, levels in injected striata were significantly (** P < .01, *** P < .001) different from levels in uninjected striata. Differences between groups with either NMDA absent or present were analyzed by ANOVA followed by Tukey-Kramer’s multiple-comparison post test.
