Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Adenosine (ADO) functions as an important homeostatic modulator of cellular function in mammalian physiology (Ralevic and Burnstock, 1998). ADO inhibits neurotransmitter release in both the central and peripheral nervous systems, providing an inhibitory buffer to excitatory neurotransmission (Williams, 1989). The effects of ADO on cellular excitability are mediated via interactions with different cell surface receptor subtypes (termed P1 receptors: A₁, A_2A, A_2B, and A₃ receptor subtypes) and can result in cellular protection during conditions of physiological stress or trauma, including ischemia, seizures, inflammation, and pain (Ralevic and Burnstock, 1998). The ability of ADO to function as an inhibitory local hormone has provided a basis for its consideration as a "retaliatory" or "homeostatic" modulator of cellular activity (Newby, 1984; Williams, 1989).

Because ADO has a half-life on the order of seconds in physiological fluids (Moser et al., 1989), its actions are generally restricted to those tissues and cellular sites where it is released. The extracellular concentration of ADO is controlled via its rapid reuptake into the cell and subsequent intracellular metabolism (Arch and Newsholme, 1978). Adenosine kinase (AK; ATP:adenosine 5'-phosphotransferase, EC 2.7.1.20) is a cytosolic enzyme that catalyzes the phosphorylation of ADO to AMP and is the predominant enzyme regulating ADO metabolism under physiological conditions (Arch and Newsholme, 1978). AK has been purified from a number of mammalian species (Pallella et al., 1980) and the cloning of two human isoforms, AK_long and AK_short, has recently been described (Spychala et al., 1996; McNally et al., 1997). There is evidence to indicate that AK may contain two ligand recognition sites, a catalytic ADO-sensitive site and a regulatory ATP-sensitive site (Hawkins and Bagnara, 1987; Lin et al., 1988). ADO levels are primarily regulated by a nonconcentrative, bidirectional, facilitated diffusion transporter (Boleti et al., 1997) and inhibition of intracellular AK has the net effect of decreasing cellular reuptake of ADO (Davies et al., 1984), thereby increasing the local concentration of ADO in the extracellular compartment.

Historically, therapeutic exploitation of the protective actions of ADO using receptor-selective agonists has been limited by unacceptable side effects associated with the activation of ADO receptors in tissues other than the therapeutic target (Williams and Jarvis, 2000). Because the actions of endogenous ADO are highly localized and AK blockade may be more effective in cells undergoing accelerated ADO release (Newby et al., 1983), inhibition of AK may result in a greater degree of therapeutic specificity (Mullane and Young, 1993). Consistent with this hypothesis, AK inhibition has been shown to increase extracellular ADO concentrations in vitro (Pak et al., 1994; Golembiowska et al., 1996) and to selectively increase brain ADO concentrations in vivo only in neural tissue undergoing trauma (Britton et al., 1999). AK inhibitors have been shown to be more effective than inhibitors of ADO deaminase in increasing the release of endogenous ADO (Golembiowska et al., 1996) and in reducing seizure susceptibility (Zhang et al., 1993) and nociception (Keil and DeLander, 1992; Poon and Sawynok, 1995) in vivo. Consequently, AK inhibition may represent a mechanism to selectively enhance the actions of ADO while minimizing nonspecific side effects associated with ADO receptor agonists (Kowaluk and Jarvis, 2000). This hypothesis is supported by recent data indicating that systemically administered AK inhibitors can reduce seizure susceptibility and anesthetic requirement at doses that do not alter cardiovascular function (Wang et al., 1997; Wiesner et al., 1999).

Demonstrations of the antinociceptive effects of AK inhibition have been primarily based on the pharmacology of intrathecally administered NH₂dADO (Fig. 1) (Keil and DeLander, 1992; Poon and Sawynok, 1995), which inhibits AK with nanomolar affinity, but has poor cell penetrability, and may have limited access to the CNS after systemic administration (Kowaluk et al., 1999). Other AK inhibitors such as 5-iodotubercidin (5-IT) and 5'-deoxy,5-iodotubercidin (5'd-5IT) (Fig. 1) have greater affinity for intracellular AK (Davies et al., 1984) and have in vivo efficacy in acute pain models (Sawynok et al., 1998; Kowaluk et al., 1999), but have poor oral bioavailability (Cottam et al., 1993; Ugarkar et al., 2000).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Structures of ABT-702, 5'd-5IT, 5-IT, and NH2dADO.

The present report describes the pharmacology of ABT-702 (Fig. 1), a novel non-nucleoside AK inhibitor that is a potent, competitive, and reversible inhibitor of AK across a variety of mammalian species. Furthermore, ABT-702 effectively reduces acute nociception in the mouse after oral administration and the analgesic actions of ABT-702 are mediated by a nonopioid, adenosine receptor-mediated mechanism.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [U-¹⁴C]Adenosine (542 mCi/mmol) and [2-³H]adenosine (26 Ci/mmol) were purchased from Amersham International (Amersham, Buckinghamshire, UK). ¹²⁵I-4-Aminobenzyl-5'-N-methylcarboxamidoadenosine (2200 Ci/mmol) and [³H]NBTI (15-30 Ci/mmol) were purchased from Amersham Inc., Arlington Heights, IL. DE-81 anion exchange filter disks were from Whatman (Maidstone, UK). [³H]Cyclohexyladenosine (27 Ci/mmol) and [³H]CGS 21680 (24 Ci/mmol) were purchased from NEN Life Science Products, Boston MA. Bovine serum albumin, ATP, ADO, and other chemical reagents were purchased from Sigma Chemical Co. (St. Louis, MO). ABT-702 was synthesized as described by Lee et al. (C.-H. Lee, M. Jiang, M. Cowart, G. Gfesser, R. Perner, K. H. Kim, Y. G. Gu, M. Williams, M. F. Jarvis, E. A. Kowaluk, A. O. Stewart, and S. S. Bhagwat, submitted for publication).

Preparation of Mammalian Brain Cytosol. Rat brains were freshly harvested from male Sprague-Dawley rats (Charles River Farms, Wilmington, MA). Whole rat brain tissue was washed and homogenized in a homogenizing buffer containing 25 mM Tris-HCl, pH 7.6; 0.1 mM EDTA; 0.1 mM EGTA; 0.1% 2-mercaptoethanol; 250 mM sucrose; 0.1 mM phenylmethylsulfonyl fluoride; 0.2 µM leupeptin; and 0.1 µM pepstatin at the ratio of 7 ml of buffer for every 9 g of brain tissue. The homogenate was centrifuged at 100,000g at 4°C for 1 h. The supernatant (cytosol) was pooled and assayed for protein concentration and adenosine kinase activity. Protein was measured by the method of Bradford (1976) with bovine serum albumin as standard. Brain tissue cytosol was also similarly prepared from monkey (cynomolgus, Macaca fascicularis; BRF, Houston, TX), dog (beagle; Marshall Farms, North Rose, NY), and mouse (CF-1; Harlan Farms, Portage, MI). All animal handling, behavioral testing, and tissue collection protocols were approved by an institutional animal care and use committee.

AK Inhibition Assay. AK activity of cell supernatants was assayed radiochemically as described by Yamada et al. (1980) with modifications. Routine enzyme inhibition assays were carried out at 23°C in a final volume of 100 µl. The reaction mixture contained 64 mM Tris-HCl, pH 7.5; 0.2 mM MgCl₂; 1 mM ATP; 0.2 µM [U-¹⁴C]adenosine or [2-³H]adenosine; and appropriate volumes of rat brain cytosol as a source of AK. The reaction was incubated for 15 min and terminated by aliquoting 40 µl of the reaction mixture onto disks of DE-81 anion exchange filter disks. The filter disks were then air dried, washed for 15 min in 2 mM ammonium formate, and then sequentially rinsed with excess distilled water, methanol, and acetone, and dried under nitrogen gas. The filter disks were then soaked for 5 min in 0.5 ml of 0.1 N HCl/0.4 M KCl before addition of 5 ml of scintillation cocktail. Bound radioactivity was determined by standard scintillation spectrometry at an efficiency of 55%.

Intact Cell ADO Phosphorylation Assay. Routine assays were conducted in 24-well tissue culture plates in a 37°C shaking water bath with a final volume of 0.5 ml. IMR-32 human neuroblastoma cells (American Type Culture Collection, Gaithersburg, MD) were grown to confluency in RPMI 1640 medium for 2 to 4 days before each experiment. Culturing media were then aspirated, 400 µl of warm Gey's balanced salt solution was added to each well, and the cells were preincubated for 10 min. Appropriate concentrations of test compounds (10¹¹-10⁴ M) were added to each well and incubated for 10 min. The reaction was initiated by the addition of 50 µl of 2 µM [U-¹⁴C]adenosine. After a 20-min incubation, the assay buffer was rapidly aspirated, 200 µl of ice-cold stop buffer (20 mM sodium acetate, pH 4.0; 2 mM EDTA) was added to each well, and the cells were quickly frozen by the addition of excess liquid nitrogen. The plates were allowed to thaw at room temperature for 20 min and a 50-µl aliquot of the supernatant was placed onto DE-81 filter disks. The filter disks were then processed as described above for the AK enzyme assay.

AK Kinetic Studies. AK assays were performed at 23°C and contained 50 mM sodium-HEPES, pH 7.5; appropriate concentrations of [U-¹⁴C]adenosine; MgCl₂; ATP; ABT-702 (0.5-2.0 nM); and rat brain cytosol as the source of AK in a total volume of 300 µl. To evaluate whether ABT-702 was competitive with ADO, experiments were performed with at a constant MgATP^2- concentration (0.24 mM MgCl₂, 0.04 mM ATP), whereas [U-¹⁴C]adenosine concentrations were varied from 0.02 to 0.12 µM. In the converse experiments, [U-¹⁴C]adenosine (0.07 µM) was kept constant and MgATP² concentrations varied from 0.015 to 0.08 mM. The reaction was initiated by addition of rat brain cytosol as the source of AK. Initial velocities were determined by withdrawing 50-µl aliquots at intervals and applying them immediately onto DE-81 filter disks. The DE-81 filters were then processed as described above for the AK assay. Reaction rates were linear for at least 15 min. Enzyme-free blank reactions were used as controls. Double-reciprocal plots of initial velocity data were analyzed taking into the consideration that MgATP², rather than ATP, is the substrate for rat brain AK, and MgATP² concentration is considered to be governed by the equilibrium MgATP²  Mg²⁺ + ATP⁴ (K_i = 0.0143 mM) (Palella et al., 1980). The results were reported as MgATP² concentration, assuming that all ATP present is complexed with magnesium when magnesium is in excess.

AK Inhibitor Reversibility Studies. The binding reversibility of ABT-702 to human recombinant AK was determined using methodology adapted from Greffard et al. (1997). ABT-702 stock solutions were made at 100 mM in dimethyl sulfoxide. Human recombinant AK was purified by affinity chromotography similar to that described for rat brain AK (McNally et al., 1997), and based on preliminary stability studies, stored at a protein concentration of 0.377 mg/ml in 20 mM Tris, 1 mM DTT, 50 mM KCl, 20% glycerol (pH 7.5). AK inhibition was carried out using 1 mM ABT-702 and time-matched controls contained equivalent volume of dimethyl sulfoxide. Both control and ABT-702-containing enzyme reactions were dialyzed using a Microdialyzer system 100 (Pierce, Indianapolis, IN) in a buffer containing 20 mM Tris-HCl, 50 mM KCl, 10% glycerol, 1 mM DTT (pH 7.5). At each time point (0.25, 0.5, 1, 2, and 4 h), 10-µl samples were taken from both control and ABT-702 dialysis reactions to assess AK activity and the dialysis buffer was changed. AK activity of all samples was assessed as described above.

ADO Selectivity Studies. Binding studies for ADO A₁, A_2A, and A₃ receptors were conducted as previously described (Jarvis, 1998). Radioligand binding studies for the rat A₁ receptor used rat cortical membranes that were incubated in an assay buffer containing 50 mM Tris-HCl (pH 7.4). The rat A_2A receptor-binding assay was conducted using rat striatal membranes under the same conditions as for the A₁ receptor with the inclusion of 10 mM MgCl₂ in the assay buffer. The A₁ receptor was labeled using 1 nM [³H]cyclohexyladenosine and the A_2A receptor was labeled using 5 nM [³H]CGS 21680. Binding reactions for both the A₁ and A_2A receptors were carried out for 2 h at 23°C. Nonspecific binding for both the A₁ and A_2A receptor-binding assays was defined in the presence of 20 µM 2-chloroadenosine.

Pharmacological Selectivity Studies. The selectivity of ABT-702 as an AK inhibitor was evaluated in a number of assays to assess pharmacological selectivity relative to other cell-surface receptors, ion channels, transport sites, and enzymes (see Table 4) by use of standardized assay protocols (Cerep, Celle l'Evescault, France).

Hot-Plate Assay. Acute thermal nociception was measured using an automated hot-plate analgesia monitor (model AHP16AN; Omnitech Electronics, Columbus, OH) using methodology as described by Kowaluk et al. (1999). Briefly, male CF-1 mice (Harlan Farms) weighing approximately 25 to 30 g were housed 14 to a cage and maintained in a climate-controlled facility with a 12-h:12-h light:dark cycle. In all experiments, drug naïve mice were used once and the temperature of the hot-plate was maintained at 55°C. All test compounds were administered i.p. at a volume of 10 ml/kg. ABT-702 or morphine was routinely administered i.p. 30 min before analgesia testing. The pretreatment time for oral administration of ABT-702 was 60 min before analgesia testing. In the antagonist studies, ADO receptor antagonists or naloxone were administered i.p. 30 min before ABT-702 or morphine, respectively. Experimental and control groups contained six to eight mice each. Mice were place in individual, 9.8 × 7.2 × 15.3-cm (l × w × h) plastic enclosures on the hot-plate and the latency until the 10th jump was recorded. Jumps were recorded by disruption of a photocell beam located 12.5 cm above the surface of the hot-plate. Mice were removed from the hot-plate after either 10 jumps were made or 180 s (test termination) had elapsed, whichever occurred first. The latency until the 10th jump was used for statistical analysis. Jump latency data were analyzed using analysis of variance. Where appropriate, Fisher's protected least significant difference was used for post hoc analysis. The level of significance was set at P < .05. ED₅₀ values were estimated using linear regression.

Abdominal Constriction Assay. The antinociceptive effects of ABT-702 were also assessed in a model of persistent chemical pain, the phenyl-p-quinone-induced abdominal constriction assay as previously described (Kowaluk et al., 1999). ABT-702 or morphine was administered i.p. to mice 30 min before receiving injections of phenyl-p-quinone (68 µmol/kg i.p. dissolved in 5% ethanol). The presence of characteristic stretching or writhing responses was noted during a 10-min period beginning 5 min after the injection of phenyl-p-quinone. Mice displaying one or more of these nociceptive responses were categorized as responders, and mice who did not display these behaviors were regarded as nonresponders. The chi square statistic was used to evaluate statistical significance (P < .05).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

In Vitro Characterization. ABT-702 potently inhibited the activity of rat brain cytosolic AK in a concentration-dependent manner with an IC₅₀ value of 1.7 nM (Table 1; Fig. 2, left). ABT-702 displayed equivalent potency to the nucleoside AK inhibitor 5'd-5IT and was approximately 9-fold more potent compared with 5-IT and NH₂dADO (Table 1). ABT-702 also potently inhibited AK activity in intact cultured IMR-32 human neuroblastoma cells (IC₅₀ = 51 nM), indicating that ABT-702 can penetrate the cell membrane and potently inhibit AK at its intracellular site (Fig. 2, right).

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Representative concentration-effect curves for ABT-702 and prototypic nucleoside AK inhibitors. Right, inhibition of AK activity from rat brain cytosol (AK enzyme assay). Left, inhibition of AK activity in intact cultured IMR-32 human neuroblastoma cells.

2

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Kinetics of AK inhibition by ABT-702. Double reciprocal plots of the rate of AK activity versus substrate concentration at different fixed concentrations of ABT-702. Top, when ADO is the varied substrate, the reciprocal plots intersect at the y-axis, indicating that ABT-702 is a competitive inhibitor of AK with respect to ADO. Bottom, when MgATP2 is the varied substrate, the concentration-dependent shift in both the slopes and intercepts (parallelism) indicates that A-200702 is a noncompetitive inhibitor of AK with respect to MgATP2. , control; , 0.5 nM A-200702; , 1.0 nM A-200702; , 1.5 nM A-200702; , 2.0 nM A-200702.

2

2

View larger version (37K): [in this window] [in a new window]   Fig. 4.   Reversible inhibition of recombinant human AKshort by ABT-702. No dialysis: control indicates mean AK in the absence of inhibitor, and ABT-702 indicates mean AK activity in the presence of excess ABT-702 (100 mM). Four-hour dialysis: control indicates mean AK activity in the absence of inhibitor after 4 h of dialysis, and ABT-702 indicates mean recovery of AK activity after 4 h of continuous dialysis with 20 mM Tris-HCl, 50 mM KCl, 10% glycerol, 1 mM DTT, pH 7.5, at 23°C. *P < .05 versus control.

Antinociceptive Effects. ABT-702 significantly reduced acute thermal nociception in a dose-dependent manner after both intraperitoneal (ED₅₀ = 8 µmol/kg i.p.) and oral (ED₅₀ = 65 µmol/kg p.o.) administration in the mouse hot-plate test (Fig. 5). In this test of acute somatic nociception, morphine also dose dependently increased jump latencies (ED₅₀ = 4 µmol/kg i.p., Fig. 5). The antinociceptive effects of ABT-702 were significantly attenuated by the nonselective ADO receptor antagonist theophylline (10 mg/kg i.p.), but not by the opioid receptor antagonist naloxone (5 mg/kg i.p.), indicating a nonopioid mechanism of action (Fig. 6). This dose of naloxone was fully effective in blocking the antinociceptive effects of morphine in this model (Fig. 6). The antinociceptive effects of ABT-702 were not blocked by the peripherally selective antagonist 8-(p-sulfophenyl)-theophylline (8-PST, 200 µmol/kg i.p.) (Fig. 6). Additional antagonist studies indicated that the ADO A₁ receptor-selective antagonist cyclopentyltheophylline (CPT, 10 mg/kg i.p.), but not the ADO A_2A receptor-selective antagonist 3,7-dimethyl-1-propargylxanthine (DMPX, 1 mg/kg i.p.) significantly attenuated the antinociceptive effects of ABT-702 (30 µmol/kg i.p., Fig. 7). Consistent with its antinociceptive effects in the hot-plate assay, ABT-702 also produced dose-dependent antinociceptive effects (ED₅₀ = 2 µmol/kg i.p.) in the abdominal constriction assay (Fig. 8). Like morphine (21 µmol/kg i.p.), ABT-702 exhibited full efficacy in this model of persistent chemical pain.

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effects of ABT-702 in the mouse hot-plate test. ABT-702 dose dependently increased the jump latencies of mice after both i.p. and p.o. administration in the hot-plate test. For comparison, morphine administered i.p. also dose dependently reduced nociception in the hot-plate test. Data are mean ± S.E.M. *P < .05 versus vehicle (n = 6-8 per dose group).

View larger version (26K): [in this window] [in a new window]   Fig. 6.   Left, antinociceptive effects of ABT-702 (30 µmol/kg i.p.) were blocked by the ADO receptor antagonist theophyllline (10 mg/kg i.p.), but not by the opioid receptor antagonist naloxone (5 mg/kg i.p.). The effects of morphine (15.4 µmol/kg i.p.) were blocked by naloxone (5 mg/kg i.p.), but not by theophylline (10 mg/kg i.p.). Right, antinociceptive effects of ABT-702 (30 µmol/kg i.p.) were blocked by the ADO receptor antagonist theophyllline (10 mg/kg i.p.), but not by the peripherally selective ADO receptor antagonist 8-PST (50 mg/kg i.p.). Data are mean ± S.E.M. *P < .05 versus respective vehicle response, +P < .05 versus corresponding response with coadministration of saline or vehicle (n = 5-8 per dose group).

View larger version (33K): [in this window] [in a new window]   Fig. 7.   Antinociceptive effects of ABT-702 (30 µmol/kg i.p.) were attenuated by the A1 receptor-selective antagonist CPT (10 mg/kg i.p.), but not by the A2A receptor-selective antagonist DMPX (1 mg/kg i.p.). Data are mean ± S.E.M. *P < .05 versus respective vehicle response, +P < .05 versus corresponding response with coadministration of vehicle (n = 5-8 per dose group).

View larger version (22K): [in this window] [in a new window]   Fig. 8.   ABT-702 dose dependently reduces nociception in the mouse abdominal constriction assay. % Responders represents the percentage of mice displaying the prototypic abdominal constriction response relative to the number tested. Test compounds were administered s.c. to individual groups of mice (n = 8) 30 min before phenyl-p-quinone. *P < .05 compared with vehicle-treated mice.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The non-nucleoside AK inhibitor ABT-702 lacks the nucleoside-like structural features common to all other reported AK inhibitors, which are structural analogs of ADO, including the low-affinity inhibitor tubercidin and the more potent AK inhibitors NH₂dADO, 5-IT, and 5'd-5IT (Kowaluk and Jarvis, 2000). These latter compounds, however, have been of limited pharmacological or therapeutic utility due to their general short half-lives in vivo, poor oral bioavailability and cell penetrability, lack of pharmacological selectivity, and their potential to form cytotoxic metabolites (Cottam et al., 1993; Wiesner et al., 1999; Ugarkar et al., 2000).

AK has been purified from a variety of tissues from different species, including humans (Pallella et al., 1980; Yamada et al., 1980). Additionally, the AK gene has been cloned and expressed from rat and human tissues (Spychala et al., 1996; McNally et al., 1997). Two isoforms of AK mRNA (AK_short and AK_long) have been found in human tissue that encode proteins that are identical in sequence except at the amino terminus, where AK_long is characterized by a 17 amino acid extension compared with AK_short (McNally et al., 1997). When expressed in Escherichia coli, both isoforms provide a soluble, active protein that exhibits the enzymatic and pharmacological characteristics of native AK (McNally et al., 1997). The present data demonstrate that ABT-702 has equivalent high potency to inhibit AK across several mammalian species, including human native AK and the human recombinant isoforms, AK_long and AK_short. ATP is generally considered to is the preferred phosphate source for the reaction catalyzed by AK and the true AK substrate is probably the MgATP² complex (Pallella et al., 1980). Two ADO binding sites have been proposed to exist on AK, a catalytic site with high affinity for ADO and a low-affinity regulatory site (Hawkins and Bagnara, 1987; Lin et al., 1988). Additional data suggest that the low-affinity ADO binding site might be the ATP binding site, and that binding of ADO to this site may be responsible for substrate inhibition by ADO (Elalaoui et al., 1987). The present kinetic studies revealed that ABT-702 is a reversible AK inhibitor that is competitive with respect to ADO and a noncompetitive inhibitor with respect to MgATP². ABT-702 also has high pharmacological specificity for inhibition of AK compared with other sites of ADO action, as well as other cell surface receptor, ion channel, enzyme, and signal transduction targets. Thus, as a novel and selective non-nucleoside AK inhibitor, ABT-702 has a molecular interaction at the AK enzyme that is similar to other potent nucleoside-like AK inhibitors (Davies et al., 1984; Ugarkar et al., 2000). The X-ray crystallographic structure of human AK has recently been described and consists of two -helix/-sheet domains, with the active site lying between the two domains (Matthews et al., 1998).

Functionally, ABT-702 demonstrated potent antinociceptive activity in the mouse hot-plate assay after both i.p. and oral administration. The antinociceptive activity of ABT-702 is consistent with the analgesic actions demonstrated for another AK inhibitor, 5'd-5IT, after systemic administration in this assay (Kowaluk et al., 1999). The analgesic potency of ABT-702 was comparable to morphine in this test and ABT-702 was more potent than other AK inhibitors such as 5IT and NH₂dADO, which do not readily penetrate into the CNS (Kowaluk et al., 1999). Consistent with its analgesic actions in the hot-plate test, ABT-702 also dose dependently reduced nociception in the mouse abdominal constriction assay, a model of persistent chemical pain. The potency of ABT-702 (ED₅₀ = 2 µmol/kg s.c.) in this test was also similar to that found for morphine (ED₅₀ = 3 µmol/kg s.c.) (Kowaluk et al., 1999).

The antinociceptive effects of ABT-702 appear to be mediated by actions at ADO receptors in the CNS because a nonselective ADO receptor antagonist, theophylline, but not a peripherally acting ADO receptor antagonist, 8-PST, significantly attenuated ABT-702-induced antinociception in the hot-plate assay. It should be noted that 8-PST is approximately 10-fold more potent than theophylline at ADO A₁ receptors (Jarvis et al., 1989), and 8-PST was used at an approximately 4-fold higher dose relative to theophylline. The dose of 8-PST has previously been shown to selectively block the peripheral effects of ADO receptor agonists (Marston et al., 1998). Both theophylline and the ADO A₁ receptor-selective antagonist CPT were found to be equally effective in attenuating the antinociceptive effects of ABT-702, indicating the that the analgesic effects of ABT-702 may be mediated by a selective interaction with ADO A₁ receptors. This idea is supported by the inability of an A_2A receptor-selective dose of DMPX (Seale et al., 1988) to significantly attenuate the analgesic effects of ABT-702 in this assay. These results are consistent with the pharmacological profile of 5'd-5IT-induced antinociception in the hot-plate assay (Kowaluk et al., 1999) and with other reports indicating that spinal ADO A₁ receptors mediate the analgesic effects of ADO (Holmgren et al., 1986; Keil and DeLander, 1992; Poon and Sawynok, 1995).

Although ABT-702 displayed equivalent potency and efficacy to morphine in reducing acute nociception in the mouse, both in vitro and in vivo data indicate that the antinociceptive effects of ABT-702 are not mediated through interactions with opioid receptors. In in vitro selectivity assays, ABT-702 showed from 2000-fold to more than 5000-fold greater affinity for AK inhibition compared with its activity at opioid receptor subtypes. Additionally, a dose of naloxone (5 mg/kg i.p.) that fully blocks the analgesic effects of morphine did not alter the antinociceptive effects of ABT-702. As noted above, the antinociceptive effects of ABT-702 could be readily attenuated by a dose of theophylline that had no effects on morphine-induced analgesia. This pharmacological profile of the antinociceptive effects of ABT-702 is consistent with other data demonstrating that the antinociceptive effects of another AK inhibitor, 5'd-5IT, are specifically mediated by ADO A₁ receptor activation (Kowaluk et al., 1999). Interestingly, the acute antinociceptive effects of 5'd-5IT and morphine have been demonstrated to be additive in the hot-plate assay (Kowaluk et al., 1999). Taken together, these data indicate that the systemic antinociceptive effects of ABT-702 and morphine are pharmacologically distinct. Although activation of opioid receptors has been reported to stimulate ADO release in vitro (Sawynok, 1997), this interaction may contribute to spinal, but not supraspinal, antinociception in vivo (Keil and DeLander, 1992; Sawynok, 1997; Lavand'homme and Eisenach, 1999).

In conclusion, the present data demonstrate that ABT-702 has high pharmacological specificity to potently inhibit AK in vitro. After systemic administration, ABT-702 produced significant acute antinociception in the mouse by a nonopioid ADO receptor-mediated action in the CNS. The ability of ABT-702 to potently and fully alleviate acute thermal nociception indicates that the localized enhancement of extracellular ADO concentrations via AK inhibition provides a novel and effective analgesic mechanism that is independent of activation of opioid receptors. AK inhibitors, including 5'd-5IT, as well as other recently described nucleoside analogs (e.g., GP683), have been shown to enhance the endogenous neuroprotective effects of ADO in both ischemia (Miller et al., 1996; Jiang et al., 1997) and seizure models (Wiesner et al., 1999) at doses that are separable from the peripheral side effects (hemodynamic and sedation) commonly associated with ADO receptor agonists (Kowaluk and Jarvis, 2000). This preclinical profile highlights the therapeutic potential of AK inhibitors in the treatment of disorders involving CNS hyperexcitability.