Introduction

Top Abstract Introduction Methods Results Discussion References

Pharmacological manipulation of adenosine receptors offers an attractive target for drug treatment of several human CNS diseases (Jacobson et al., 1992). For example, adenosine receptor antagonists may have potential as cognitive enhancers for use in chronic neurodegenerative disorders, a hypothesis supported by in vivo pharmacological evidence for a role of adenosine in cognitive function (Normile and Barraco, 1991; Schingitz et al., 1991; Hitchcock et al., 1992; Von Lubitz et al., 1993; Suzuki et al., 1993; Normile et al., 1994; Jacobson et al., 1996). Adenosine receptor ligands also reduce cell death in animal models of cerebral ischemia (Jacobson et al., 1996). The issue of blood-brain barrier penetration is an important consideration if therapeutically useful compounds are to be developed, requiring the development of simple models for predicting brain penetration. For CNS drugs, a correlation between expression of a centrally mediated behavior and the brain concentration of drug, after peripheral administration can provide a valuable first indication of clinical potential. Spontaneous locomotor activity is one example of a simple, centrally mediated behavior amenable to pharmacological manipulation. Further, a hypolocomotor response is common to the effects produced by selective adenosine A₁, A_2A and A₃ receptor agonists (Barraco et al., 1983; Phillis et al., 1986; Heffner et al., 1989; Durcan and Morgan, 1989a, b; Nikodijevic et al., 1990, 1991; Brown et al., 1991; Jacobson et al., 1993b; Zarindast and Heidari, 1993). Conversely, administration of adenosine A_2A receptor antagonists induces a mild hyperlocomotor response, whereas adenosine A₁ receptor antagonists do not effect spontaneous locomotor activity (Nikodijevic et al., 1991; Griebel et al., 1991; Holtzman, 1991; Jacobson et al., 1993a). However, it is presently unclear whether this behavioral profile is mediated by dependent or independent mechanisms.

The study of the behavioral sequelae of adenosine receptor modulation has been greatly assisted by the recent development of receptor selective drugs. The pharmacological subdivision of adenosine receptors into four subtypes, termed A₁, A_2A, A_2B and A₃ receptors, has been confirmed by molecular cloning studies (Fredholm et al., 1994). Adenosine analogues such as CPA (Williams et al., 1986), APEC (Nikodijevic et al., 1991) and N⁶-(3-iodobenzyl)-5'-methylcarboxamidoadenosine (Jacobson et al., 1993b) are selective agonists for adenosine A₁, A_2A and A₃ receptors, respectively. Specific adenosine A₁ receptor antagonists include both xanthine derivatives such as DPCPX (Bruns et al., 1987) and the novel pyrazolopyridines, FK453 and FK352 (Terai et al., 1995; Maemoto et al., 1995; Ito et al., 1995; Maemoto et al., 1997). Selective adenosine A_2A receptor antagonists, including KF17837 (Nonaka et al., 1994), have also been synthesized recently (Ongini and Fredholm, 1996), although 3-(3-iodo-4-aminobenzyl)-8-(4-oxyacetate)-1-propylxanthine inhibits peripheral adenosine A₃ receptor responses (Fozard and Hannon, 1993).

Our aims were to characterize the pharmacological profile of adenosine receptor mediated alterations in spontaneous locomotion in rats, and to determine whether this model provides a useful model for predicting blood brain barrier penetration of adenosine A₁ receptor antagonists. This approach used a locomotor activity paradigm developed by this group (Marston et al., 1994), combined with measurements of antagonist concentration in brain and serum using a modified radioreceptor assay (Finlayson et al., 1997), and in vitro determination of drug affinity using a radioligand binding assay (Maemoto et al., 1997). The resulting data proved sufficiently reliable to allow the mathematical relationship between equivalent behavioral dose, drug concentration, and in vitro receptor affinity to be explored.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animals

Twenty-two groups of eight male Sprague-Dawley rats (Charles Rivers, Margate, UK) were used, each to test only one agonist, or one agonist/antagonist combination. Animals were group housed under a 12-hr light/dark cycle in a temperature-controlled environment (21 ± 1°C). At the start of the experiment, animals were between 220 and 250 g in weight. A mild food restricted feeding regimen was used throughout the experiment, which allowed the animals to gain 2 to 4 g body weight per week. Water was available ad libitum throughout the studies.

Locomotor Activity

Spontaneous locomotor activity was monitored using a double infrared beam system. Eight clear polycarbonate cages (480 mm long × 270 mm wide × 200 mm high) were mounted on a modified holding cage rack (RS Biotech, Alva, Scotland). Each cage was fitted with a flush flat wire lid and floor, and beneath each cage was positioned a waste tray thinly cover with a layer of dustless sawdust. Two infrared beams were positioned across each cage, 30 mm above the floor of the cage and 100 mm from each end. Each beam was connected to a detection system, which registered beam interruptions. This information was in turn passed to a BeeBex computer interface (Paul Fray Ltd, Cambridge, England). An Acorn A5000 computer was used to collect, interpret, collate and store the data received from the detection system. The software required was written in house using the Arachnid extension to BBC Basic (Paul Fray Ltd.). An activity count was recorded, in the appropriate time-bin, only when both beams were broken sequentially.

Experimental design. The dose relationship for adenosine receptor agonist-induced alterations in locomotor activity was examined. These experiments were carried out according to the same general timetable. A drug test day was always preceded by a vehicle test day and followed by a no-drug, or washout, day. If at any point the level of activity on a vehicle day was out of the expected range, vehicle days were repeated until the baseline was restabilized. Drug doses were administered in random sequence with administration 5 min before a 30 min locomotor test session.

1

1

View this table: [in this window] [in a new window]   TABLE 1 Identity, full chemical names, vehicle and dose range (µmol kg1) of adenosine receptor agonists and antagonists

Statistical Analysis of Behavioral Data

One-way analysis of variance (SigmaStat, Jandel Scientific, Erkrath, Germany) was used to determine drug effects on spontaneous locomotor activity. In all cases, normality and equality of variance were assessed before performing parametric analysis. When appropriate a logarithmic transformation was applied to the locomotor counts when the spread of variance was found heterogeneous. After analysis of variance, post hoc tests were carried out using Tukey's procedure to determine significant differences from either antagonist alone or vehicle as control. An EBD was derived as the lowest dose of antagonist to reverse hypolocomotion to a point not significantly different from vehicle baseline. For comparison of drug potency ED₅₀, or ID₅₀, values were determined by selecting the three doses from the steepest component of the dose response curve. Least squares linear regression was used to determine the gradient of the line and the derived equation used to determine a point mid-way between the vehicle and agonist alone baseline.

Radioreceptor Assay

In vivo procedures and sample preparation. Male Sprague Dawley rats (280-300 g; Charles River) were injected i.p. with vehicle or drug at the equivalent behavioral dose (table 2). Eighteen min afterinjection, animals were anaesthetized with 4% halothane in oxygen and nitrous oxide (30:70 v/v), and venous blood (7 ml) was collected from the inferior vena cava. Blood was allowed to clot at 4°C overnight, and the resulting serum stored at -20°C in 200 µl aliquots. Unless stated otherwise, animals were perfused at 20 min postinjection with saline (30 sec; flow rate = 35 ml min¹) via a needle inserted into the aorta through the left ventricle (Finlayson et al., 1997). After perfusion, the brain was immediately removed, the cortex dissected, rolled on filter paper to remove superficial blood vessels, weighed and homogenized in 9 vol (v/w) of 50 mM Tris-HCl (pH 7.4) (Tris-buffer) using a glass/Teflon homogenizer. An aliquot was removed for experiments using fresh brain homogenate, with the remainder stored in 1 ml aliquots at -20°C. Denaturation of brain homogenates prepared from vehicle and drug treated animals by treatment at 80°C for 15 min was used to abolish [³H]DPCPX binding capacity (Finlayson et al., 1997). For consistency, serum samples (diluted 10-fold in Tris-buffer) were also denatured. After denaturation, samples were allowed to cool and then incubated for 1 hr at 37°C with ADA (6 IU ml¹) to remove endogenous adenosine present in the sample, before determination of drug concentrations using a conventional [³H]DPCPX binding assay (Finlayson et al., 1997).

View this table: [in this window] [in a new window]   TABLE 2 Brain and serum concentrations (nM, mean ± S.E.M.) for nonstimulant adenosine receptor antagonists detected 20 min after administration of the EBD (µmol kg1)

P₂ membrane preparation. Crude synaptosomal P₂ membranes were prepared from the cerebral cortex of untreated rats. Rats were killed by decapitation and the cortex was homogenized using a glass/Teflon homogenizer in 15 vol (v/w) of ice cold 0.32 M sucrose. The homogenate was centrifuged at 1000 × g for 10 min at 4°C, and the supernatant centrifuged at 17,000 × g for 20 min at 4°C. The resulting P₂ pellet was lysed in 30 volumes of ice cold glass distilled water for 30 min. Membranes were then centrifuged at 50,000 × g for 10 min at 4°C, and the pellet resuspended in 30 volumes of Tris-buffer. After centrifugation at 50,000 × g for 10 min at 4°C, the pellet was resuspended in 5 volumes of Tris-buffer and stored in 1.4-ml aliquots at -20°C. On the day of the assay, membranes were resuspended in 30 volumes of Tris-buffer and centrifuged at 50,000 × g for 10 min at 4°C. The final pellet was resuspended in 200 volumes of Tris-buffer and stored on ice until required.

Binding assay. [³H]DPCPX (98.1 Ci mmol¹) binding was performed at equilibrium by preincubating 10 µl of DMSO or competing drug, 240 µl of Tris-buffer, 100 µl of 1 IU ml¹ adenosine deaminase and 100 µl of 1 nM [³H]DPCPX with 500 µl of the P₂ rat cortical membrane suspension (20-40 µg) and 50 µl of the denatured brain or serum sample for 20 min at 25°C (Finlayson et al., 1997). Competing drugs were diluted in DMSO to give 10 duplicate concentrations. Total binding was determined in the presence 1% DMSO, and 10 µM R-PIA was used to determine nonspecific binding. Preliminary experiments demonstrated that addition of denatured brain or serum samples did not effect the binding site affinity of any drug tested, and that affinity was unaltered by treating the drugs at 80°C for 15 min to mimic the denaturation step (data not shown). The assay was terminated by rapid filtration over GF/B filters, followed by three washes (3 ml) with Tris-buffer using a Brandel cell harvester. Filter disks were transferred to scintillation vials, incubated with 100 µl of 100% formic acid for 10 min before addition of 4 ml Emulsifier Safe scintillant and overnight equilibration. Radioactivity was determined using a Packard TR2500 liquid scintillation counter with automatic quench correction. Protein concentration was measured according to the method of Bradford (1976).

Analysis of Binding Data

Concentration response curves for competing drugs in the [³H]DPCPX binding assay were obtained in the presence of denatured brain homogenate or serum from vehicle-treated rats. The drug concentration in brain homogenate or serum samples from drug-treated animals was determined by comparing the percent inhibition of [³H]DPCPX binding in the presence of test sample with the appropriate concentration response curve. Drug concentrations in test samples were determined in duplicate using at least three animals in each treatment group, and allowing for sample dilution, estimates of drug concentration in brain and serum were obtained.

Materials

[³H]DPCPX was purchased from New England Nuclear, Stevenage, UK. DPCPX, DPSPX, caffeine, theophylline, NECA, CADO, CPA, CHA, CPT, DPX, 8-PT, R-PIA and 8-PST were purchased from Research Biochemicals Incorporated (RBI), Natick, MA. FK453, FK352, KF17837, FR160492 (patented by Takeda Pharmaceutical Co., Osaka, Japan), KW3902 and MDL102234 were synthesized by the New Drug Research Laboratories, Fujisawa Pharmaceutical Co. Ltd., Osaka, Japan. APEC was provided by Research Biochemicals International as part of the Chemical Synthesis Program of the National Institute of Mental Health, Rockville, MD, contract N01 MH30003. All other chemicals were obtained from Sigma, Fisons, Loughborough, UK and RBI and were of the highest grade available.

	Results

Top Abstract Introduction Methods Results Discussion References

Behavior

Experiment 1: agonists. Adenosine receptor agonists produced reliable, dose dependent suppression of spontaneous locomotor activity (fig. 1). Analysis of variance confirmed that each agonist significantly reduced locomotor activity compared with the corresponding vehicle baselines; NECA [F_(4,35) = 24.27, P < .001]; APEC [F_(4,35) = 46.26, P < .001], CHA [F_(4,35) = 18.18, P < .001], R-PIA [F_(4,35) = 26.26, P < .001], CPA [F_(5,42)=25.54, P < .001], CADO [F_(4,35) = 20.73, P < .001]. Based on ED₅₀ values (table 3), the rank order of potency was NECA > APEC = CHA > CPA = R-PIA > CADO. The minimum effective hypolocomotor dose of the adenosine A₁ selective agonist, CPA (0.19 µmol kg¹) and the A_2A selective agonist, APEC (0.10 µmol kg¹) was used to examine antagonist effects (P < .01, Tukey's test).

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Effects of adenosine receptor agonists on spontaneous locomotor activity. All drugs significantly (P < .01) depressed the level of locomotor activity with reference to vehicle baseline (refer to text for statistics). Data expressed as mean (n = 8) ± S.E.M., vehicle baseline represented by dotted lines plotted at ± 1 S.E.M.

View this table: [in this window] [in a new window]   TABLE 3 Adenosine receptor agonist ED50 values (µmol kg1) for the depression of spontaneous locomotor activity

Experiment 2: antagonists

Preliminary studies examined a wide dose range for each antagonist to identify any deleterious side effects (data not shown). No qualitative gross behavioral alterations were noted.

Xanthines. The xanthine based adenosine receptor antagonists; DPCPX, 8-PT, DPX, MDL102234 and KW3902, did not effect locomotor activity at the highest dose administered (fig. 2, left panel; see table 1 for doses). However, each drug reversed the locomotor depression produced by CPA (0.19 µmol kg¹) in a dose-related manner, to a level statistically indistinguishable from the vehicle treated baseline (fig. 2, right panel) (P > .05), as confirmed by analysis of variance; KW3902 [F_(5,42) = 23.2, P < .001], DPCPX [F_(5,42) = 15.2, P < .001], DPX [F_(3,28) = 14.1, P < .001], MDL102234 [F_(5,42) = 12.5, P < .001], 8-PT [F_(5,42) =11.5, P < .001]. On the basis of ID₅₀ values the rank order of potency was: KW3902 > DPX = DPCPX > MDL102234 > 8-PT (table 4). The EBD (table 2) was identified using post hoc analysis (P > .01) after analysis of variance.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   Attenuation of CPA-induced hypolocomotion by xanthine-based adenosine receptor antagonists. Left panel illustrates the effects of the drugs administered alone at the highest dose used, plus vehicle and CPA (0.19 µmol kg1). Right panel plots attenuation of CPA-induced hypolocomotion against antagonist dose. For details of the significant reversal by each drug refer to text. All data plotted as mean (n = 8) ± S.E.M., dotted lines correspond to range of vehicle baseline data.

View this table: [in this window] [in a new window]   TABLE 4 ID50 (µmol kg1) values for attenuation of CPA-induced hypolocomotion by adenosine receptor antagonists

Nonxanthines. The pyrazolopyridine based adenosine receptor antagonists; FK453, FK352 and the pyrazolopyrimidine, FR160492, did not stimulate locomotor activity when administered alone at the highest dose tested (fig. 3, left panel; see table 1 for doses). Each antagonist reversed CPA-induced (0.19 µmol kg¹) hypolocomotion in a dose-dependant manner, to a level indistinguishable from the vehicle-treated baseline (fig. 3, right panel), as confirmed by analysis of variance; FK352 [F_(3,28) = 9.5, P < .001], FK453 [F_(6,49) = 11.4, P < .001], FR160492 [F_(5,42) = 11.2, P < .001]. On the basis of ID₅₀ values the rank order of potency was: FK352 > FK453 > FR160492 (table 4). The EBD (table 2) was identified using post hoc analysis (P > .01) after analysis of variance.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Attenuation of CPA-induced hypolocomotion by nonxanthine-based adenosine receptor antagonists. Left panel illustrates the effects of the drugs administered alone at the highest dose used, plus vehicle and CPA (0.19 µmol kg1). Right panel plots attenuation of CPA-induced hypolocomotion against antagonist dose. Data for the reference compound DPCPX reproduced from figure 2 for reference. For details of the significant reversal by each drug refer to text. All data plotted as mean (n = 8) ± S.E.M., dotted lines correspond to range of vehicle baseline data.

Xanthine-based stimulants. Four adenosine receptor antagonists; caffeine, theophylline, CPT and KF17837, stimulated locomotor activity above the vehicle baseline when administered alone at the highest dose tested (fig. 4, left panel; see table 1 for doses). Theophylline, caffeine and CPT also reversed the locomotor depressive effect of CPA (0.19 µmol kg¹), and at higher doses significantly elevated locomotor activity above baseline (fig. 4, right panel). Theophylline attenuated the CPA effect at 5.5 µmol kg¹, while 55.0 µmol kg¹ significantly elevated activity above the vehicle baseline [F_(4,35) = 71.4, P = .003]. Caffeine also attenuated the CPA-induced hypolocomotion at 5.15 µmol kg¹, but 51.5 µmol kg¹ increased locomotor activity above vehicle baseline [F_(5,42) = 32.5, P < .001]. Similarly, CPT at 12.1 µmol kg¹ stimulated activity over the vehicle baseline, whereas 1.21 µmol kg¹ attenuated CPA-induced hypolocomotion [F_(4,35) = 34.4, P < .001]. KF17837 stimulated locomotor activity when administered alone at 7.31 µmol kg¹ [F_(3,28) = 23.1, P < .001], but did not significantly reverse the hypolocomotor effect of CPA at any dose tested [F_(3,28) = 32.2, P < .001].

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Attenuation of CPA-induced hypolocomotion by stimulant adenosine receptor antagonists. Left panel illustrates the stimulant effects of the drugs administered alone at the highest dose used, plus vehicle and CPA (0.19 µmol kg1). Right panel plots attenuation of the CPA-induced hypolocomotion against antagonist dose. Data for the reference compound DPCPX reproduced from figure 2 for reference. For details of the significant reversal and stimulation by each drug, except KF17837, refer to text. All data plotted as mean (n = 8) ± S.E.M., dotted lines correspond to range of vehicle baseline data.

Peripherally active antagonists. The peripherally active adenosine receptor antagonists, DPSPX and 8-PST, did not effect spontaneous locomotor activity when administered alone, or reverse the hypolocomotor effect of CPA (0.19 µmol kg¹), when administered at the highest practicable dose (DPSPX - 14.3 µmol kg¹, 8-PST - 59.5 µmol kg¹) [F_(6,49) = 53.9, P < .001] (fig. 5). Post hoc analysis confirmed that neither DPSPX nor 8-PST attenuated CPA-induced hypolocomotion with respect to vehicle baseline (P > .05).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Failure to attenuate CPA-induced hypolocomotion by peripherally active adenosine receptor antagonists. Left panel illustrates the effects of the drugs administered alone at the highest dose used, plus vehicle and CPA (0.19 µmol kg1). Right panel plots drug effect on CPA-induced hypolocomotion (refer to text for statistics). All data plotted as mean (n = 8) ± S.E.M.

Experiment 3: APEC-induced hypolocomotion. The APEC-induced (0.10 µmol kg¹) depression of spontaneous locomotor activity was fully reversed by 7.31 µmol kg¹ KF17837 [F_(4,35) = 22.6, P < .001] (fig. 6, right panel). However, KF17837 did not stimulate activity above the vehicle baseline in the presence of APEC, unlike when administered alone [F_(4,35) = 44.5, P < .001] (fig. 6, left panel). DPCPX did not reverse APEC-induced hypolocomotion [F_(2,28) = 31.3, P < .001] at doses in excess of those capable of fully reversing CPA-induced hypolocomotion (fig. 2).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Attenuation of APEC-induced hypolocomotion by adenosine receptor antagonists. Left panel illustrates the effects of the drugs administered alone at the highest dose used, plus vehicle and APEC (0.10 µmol kg1). Right panel plots attenuation of APEC-induced hypolocomotion against antagonist dose. For details of the significant reversal by KF17837, but not DPCPX, refer to text. All data plotted as mean (n = 8) ± S.E.M., dotted lines correspond to range of vehicle baseline data.

Brain and Serum Drug Concentration

Brain and serum concentrations of drug were determined 20 min after i.p. injection of an EBD using a modified radioreceptor assay. The xanthine based adenosine receptor antagonists; DPCPX, DPX, 8-PT and MDL102234 exhibited similar brain to serum concentration ratios of 0.66 to 1.22 (table 2). KW3902 was detected in brain tissue but not in serum. The nonxanthine adenosine receptor antagonists, FK453 and FK352, were detected in both brain and serum. In contrast to the xanthine-based antagonists, these pyrazolopyridine derivatives were detected at 3.5- to 11.9-fold higher concentrations in brain than in blood 20 min after systemic dosing at the EBD. The pyrazolopyrimidine derivative, FR160492 was only detected in brain 20 min after i.p. injection of the EBD.

Regression Analyses

Regression analyses was used to explore whether a mathematical relationship exists between the affinity (K_i) of nonstimulant adenosine receptor antagonists for [³H]DPCPX binding sites (an adenosine A₁ receptor selective radioligand; Bruns et al., 1987), and 1) brain concentration detected 20 min after i.p. injection of the EBD, 2) serum concentration at the same time-point and 3) EBD. The least squares linear regression models plotted in figure 7 on log transformed data were all significant (P < .05). They confirmed a strong correlation between K_i and brain concentration (r² = 0.95), with weaker correlations between K_i and serum concentration (r² = 0.65) and K_i and EBD (r² = 0.32).

View larger version (9K): [in this window] [in a new window]   Fig. 7.   Correlation of equivalent behavioral dose (EBD, µmol kg1; lower panel), serum concentration (nM; center panel) and brain concentration (nM; upper panel) of the various adenosine receptor antagonists with Ki (nM) derived from [3H]DPCPX binding assays. In each case the r2 value of the correlation is quoted.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Our study demonstrates a dose dependent suppression of spontaneous locomotor activity in rats after systemic administration of six adenosine receptor agonists. Qualitative inspection revealed no gross differences in behavioral repertoires, although at the highest doses tested each agonist induced transient mild catalepsy and the animals were cold to the touch. These observations are consistent with previous reports that adenosine receptor agonists induce both hypolocomotion and hyperthermia in mice (Barraco et al., 1983; Heffner et al., 1989; Nikodijevic et al., 1990, 1991; Zarindast and Heidari, 1993). Both CPA and APEC, selective agonists for adenosine A₁ and A_2A receptors, respectively (Williams et al., 1986; Nikodijevic et al., 1991), induced hypolocomotion. This suggests that adenosine A₁/A_2A receptor selectivity is not a simple determinant of locomotor activity, an hypothesis supported by the antagonist data described below. CPA-induced hypolocomotion was not attenuated by the established, 8-PST (Daly, 1982; Baumgold, et al., 1992), or a putative, DPSPX, peripheral adenosine receptor antagonists. Both drugs were administered at the highest practicable doses, the former within the dose range found to antagonize the cardiovascular effects of adenosine (Evoniuk, et al., 1987). This result confirms that adenosine A₁-receptor-mediated locomotor depression is mediated centrally.

Although the psychostimulant actions of methylxanthines in mice are well documented (Snyder et al., 1981; Nikodijevic et al., 1993), the recent availability of receptor selective antagonists has prompted further pharmacological studies. When administered alone, the adenosine A_2A receptor antagonist, 8-(3-chlorostyryl)-1,3,7-trimethylxanthine is reported to be mildly stimulant (Jacobson et al., 1993a). By contrast, adenosine A₁ receptor antagonists, such as DPCPX, do not influence locomotor activity in mice (Nikodijevic et al., 1991; Griebel et al., 1991; Brockwell and Beninger 1996). This latter observation raises the question of whether adenosine A₁ receptor antagonists have an identifiable, unique central role. Our study confirms that caffeine, theophylline and CPT stimulate locomotor activity in rats, and demonstrates that the selective adenosine A_2A antagonist, KF17837 is a psychomotor stimulant (Nonaka et al., 1994). While the pharmacological substrate of methylxanthine-stimulated hyperlocomotion has still to be defined, antagonism of adenosine A_2A receptors provides the most likely explanation. A possible exception to this hypothesis would appear to be the stimulant profile of CPT, a selective adenosine A₁ receptor antagonist based on in vitro binding studies (Maemoto et al., 1997). However, a very rapid brain entry of CPT has been demonstrated (Baumgold et al., 1992), and high transient drug levels may inhibit brain adenosine A_2A receptors in vivo, regardless of the apparent in vitro selectivity of this drug. In contrast to the methylxanthines derivatives described above, this study demonstrates that the majority of adenosine A₁ receptor antagonists, including DPCPX, FK453 and FK352, do not effect spontaneous locomotor activity in rats. Although some compounds might have proved stimulant at higher doses, this is unlikely with respect to DPCPX, MDL102234 and FK453, as all demonstrated a clear response plateau. For the remaining drugs it proved impossible to administer higher doses without using vehicles that proved to have significant behavioral sequelae.

Our study provides a clear pharmacological dissociation of adenosine receptor mediated behaviors by demonstrating a dose dependent reversal of agonist-induced hypolocomotion by receptor selective antagonists. CPA-induced hypolocomotion was inhibited by the adenosine A₁ antagonist DPCPX, but not the A_2A antagonist KF17837, which exhibits 35-fold selectivity for the adenosine A_2A receptor; affinity for [³H]CGS21680 binding sites was 2.32 nM, compared with 80.6 nM in a [³H]DPCPX binding assay (Finlayson K, unpublished observations). The behavioral profile is therefore comparable to that reported by Dionisotti et al. (1994) using rat atria. The hypolocomotor effect of CPA was also reversed by a range of antagonists including the xanthines, DPX, MDL102234, 8-PT and KW3902, and the nonxanthines FK453, FK352 and FR160492. In contrast, KF17837 but not DPCPX reversed the hypolocomotor effect of APEC. Binding studies confirmed that APEC exhibits 110-fold selectivity for the adenosine A_2A receptor; APEC affinity for [³H]CGS21680 binding sites was 5.57 nM, compared with 602 nM in a [³H]DPCPX binding assay (Finlayson K, unpublished observations). These data are consistent with the view that the psychomotor sequelae of adenosine A₁ and A_2A receptor activation are mediated through different and independent substrates. This hypothesis differs from the interdependent hypothesis proposed recently (Brockwell and Beninger, 1996).

Having concluded that adenosine A₁ and A_2A receptors contribute independently to the control of locomotor activity, the relationship between the behavioral profiles of individual antagonists and adenosine A₁ receptor occupancy was explored, the latter variable extrapolated from drug concentrations determined in the radioreceptor assay and [³H]DPCPX binding site affinity (K_i) (Maemoto et al., 1997). If the behavioral sequelae are mediated by central adenosine A₁ receptors, a strong correlation between brain concentration and K_i might be anticipated with weaker associations against serum concentration and EBD. The poor correlation between the EBD and [³H]DPCPX binding site affinity indicates that the dose relationship for antagonists is not directly related to an effect of molar concentration operating under zero order kinetics in a single compartment system. Similarly, it is unlikely that the principal substrate mediating the behavioral response is located peripherally. First, the peripheral adenosine receptor antagonists, DPSPX and 8-PST, were not detected in brain tissue, despite micromolar concentrations in serum (Finlayson et al., 1997; also see Baumgold et al., 1992). This finding is consistent with their inability to reverse CPA-induced hypolocomotion. Second, using log-transformed data, the correlation between serum concentration and receptor affinity was only moderate compared to the excellent correlation between K_i and brain concentration. Taken together these findings support the hypothesis that a central mechanism mediates the reversal of CPA-induced hypolocomotion by adenosine A₁ receptor antagonists. The estimated brain concentration of antagonists was 50- to 500-fold higher than corresponding K_i values determined in [³H]DPCPX binding studies (Maemoto et al., 1997), if free distribution of drug, and an uncomplicated in vivo relationship between drug concentration in brain and receptor occupancy are assumed. Simple agonist/antagonist competition for occupancy of adenosine A₁ receptors is therefore likely to account for the observed behavioral dose response profiles.

In conclusion, this study illustrates that adenosine A₁ and A_2A receptors mediate identifiable and independent central behavioral responses. The finding that reversal of CPA-induced hypolocomotion by adenosine A₁ receptor antagonists is related to the brain concentration of drug confirms the hypothesis that receptor occupancy is a prerequisite for behavioral efficacy. This finding validates the use of this simple in vivo assay as a method for predicting the blood brain barrier permeability of adenosine A₁ receptor antagonists.