Supraspinal Antinociceptive Effects of μ and δ Agonists Involve Modulation of Adenosine Uptake

Approval for this study was obtained from the Animal Care and Use Committee (Université de la Méditerranée, CHU Timone, Marseille, France). C57BL/6 female mice (weight, 25–30 g) bred in our laboratory were used at the age of 8 weeks. Animals were housed 10 per cage and had free access to food and water in a controlled environment. The room was maintained at 21–23°C on a 12-h light–dark cycle. Each animal was tested only once. The experimenter was blinded to the pharmacologic treatment.

Adenosine (crystallized, 99% pure) and dipyridamole (5 mg/ml) were obtained from Boehringer Mannheim (Meylan, France). Methylene-adenosine-5′-diphosphate (AOPCP), nitrobenzylthioinosine, adenosine deaminase, and deoxycoformycin were obtained from Sigma (St. Quentin Fallavier, France). 9-Erythro (2-hydroxy-3 nonyl) adenine (EHNA) was purchased from ICN Pharmaceutical (Orsay, France). The reverse-phase chromatography column, methanol, and other reagents were obtained from Merck (Darmstadt, Germany). 2 Chloro-N6-cyclopentyladenosine (CCPA), 2-p-(2-carboxyethyl)-phenethylamino-5′-N-ethylcarboxamidoadenosine (CGS21680), oxymorphone hydrochloride, and caffeine (Sigma) were dissolved in 30% methanol at a concentration of 10 mm. (+)-(5α, 7α, 8β (−)-N-methyl-N (7-(1-pyrrolidinyl) 1-oxaspiro (4.5)dec-8-yl) benzenacetamide (U69593; Sigma, RBI) was dissolved in 0.1 N HCl to a concentration of 100 mm and then diluted with 30% methanol to a final concentration of 10 mm. (+)-4-[(αR)-α-((2S,5R)-4-allyl-2,5-dimethyl-1-piperazinyl)-3-methoxy-benzyl]-N,N-diethylbenzamide) (SNC80; Sigma) was dissolved in 10 mm dimethyl sulfoxide (Sigma) and then diluted with 30% methanol to a final concentration of 10 mm.

Intracerebroventricular injections were performed under isoflurane anesthesia (Forene®; Abbott, Rungis, France).

Mice (n = 8 per group) were injected intracerebroventricularly with CCPA, CGS21680, oxymorphone, SNC80, or U69593. Drugs (except caffeine) were injected alone at a concentration of 1, 10, or 50 nmol in 5 μl and in combination at a concentration of 10 nmol. Caffeine was injected at 1 nmol alone or in association.

Animals were placed on a hot plate heated to 51 ± 0.5°C (Hot-Plate analgesia meter; Harvard Apparatus, Holliston, MA), and the length of the latency period, from the placing of the animal on the hot plate until licking of the front paw, was measured 5, 15, and 30 min after the intracerebroventricular injection.

Nociceptive response was assessed by the tail-pinch test, as previously described. 12In brief, mice were first tested by pinching the tail base with an artery clip (3 mm width, 500 g constant pressure); only mice displaying a nociceptive response (biting the clip or vocalizing) within 3 s were used for the experiment. We used a cutoff time of 16 s to prevent tissue damage. The nociceptive response to tail pinching was evaluated 5, 15, and 30 min after intracerebroventricular injection.

Twelve mice were decapitated, and their brains were quickly removed. The brains were immediately homogenized in Tris buffer (50 mm Tris, 10 mm MgCl², pH 7.4; 3 ml per gram of tissue), and the homogenates were centrifuged at 1,200 g  for 10 min at 4°C. The supernatants were then centrifuged at 9,000 g  for 60 min at 4°C. The pellets were resuspended in 1.5 ml Tris buffer containing 2 U adenosine deaminase, 100 UI/ml, and incubated for 30 min at 37°C. The protein content of preparations was determined in microtiter plates, using the Micro BCA protein assay reagent (Pierce, Montluçon, France) with bovine serum albumin as the standard.

The methods used to determine adenosine uptake have been described elsewhere. 13Briefly, 125-μl samples of brain extracts (six samples each time) were kept on ice and mixed with 125 μl NaCl (0.9%) and 2 μm CaCl², to which U69593, oxymorphone, or SNC80 was added to a final concentration of 1, 10, or 100 μm. The samples were incubated for 10 min at 37°C and thoroughly mixed with the Vortex® system (Bioblock, Strasbourg, France). We then added 2 nmol adenosine (100 μl). The adenosine degradation and uptake reactions were stopped 5, 30, and 60 s after adding adenosine by adding 750 μl cold stopping solution (2 mm dipyridamole, 100 nm nitrobenzylthioinosine, 4.2 mm Na²EDTA, 5 mm EHNA, 79 mm AOPCP, and 1 μm deoxycoformycin in 0.9% NaCl). Dipyridamole and nitrobenzylthioinosine were added to stop the uptake and release of adenosine by brain synaptosomes; AOPCP was added to prevent adenosine formation catalyzed by a 5′ nucleotidase. Deoxycoformycin was used to prevent adenosine deaminase activity. Samples were centrifuged (0°C, 1,500 g , 10 min), freeze-dried, and stored at −80°C before being used for chromatography.

The technical procedure used has been described elsewhere. 14,15Briefly, a Hewlett Packard HP 1100 (Les Ulis, France) modular system was used, with a diode array detector (G13135A) and a deuterium lamp (slit 8 nm). The column (150 × 4 mm) was packed with RP8, and the injection loop volume was 200 μl. The column was equilibrated with 50 mm phosphate buffer, pH 4. Samples (100 μl) were mixed with 100 μl phosphate buffer, 100 mm (NaH²PO⁴–Na²HPO⁴; pH 4), and eluted with a methanol gradient (3 min with 0% methanol, followed by a gradient of 10–25% methanol over 15 min, returning to 0% methanol in 2 min). The intraassay and interassay coefficients of variation were between 0.5 and 1%, and the detection limit for adenosine at 254 nm was 1 pmol in an injected volume of 200 μl. During chromatography, the spectra of peaks eluted between 12 and 15 min were recorded automatically at the rate of 6 spectra per second, in the 190- to 400-nm window, at 2-nm intervals. Compounds were identified on the basis of elution time and by comparing spectra with internal references. Quantification was achieved by comparing the areas obtained with those obtained with known quantities of exogenous adenosine.

Two-way analysis of variance was used to compare groups in the in vivo  study and to compare supernatant adenosine concentrations in the in vitro  study. The Wilcoxon test was used to compare latencies over time, with each animal used as its own control. A P  value of less than 0.05 was considered significant.

At 1 nmol, none of the drugs tested had significant effects on tail-pinch test and hot plate test latencies. At 10 nmol, oxymorphone, SNC80, and U69593 increased tail-pinch test latencies by 170, 190, and 185%, respectively, compared with 1 nmol (mean of the three times, P < 0.01;fig. 1A) and increased hot plate test latencies by 136, 100, and 51%, respectively (P < 0.01;fig. 1B). At 50 nmol, oxymorphone, SNC80, and U69593 increased tail-pinch test latencies by 110, 185, and 93%, compared with 10 nmol (P < 0.05;fig. 1A) and increased hot plate latencies by 50, 51, and 72%, respectively, compared with 10 nmol (P < 0.05;fig. 1B).

At 10 nmol, CCPA and CGS21680 increased tail-pinch test latencies by 200 and 130%, respectively, compared with 1 nmol (P < 0.01) and increased hot plate latencies by 95 and 52%, respectively, compared with 1 nmol (P < 0. 01). At 50 nmol, CCPA and CGS21680 increased tail-pinch test latencies by 139 and 124%, respectively, compared with 10 nmol and increased hot plate test latencies by 118 and 65%, respectively, compared with 10 nmol (P < 0. 05)

No increase in latency was observed 30 min after intracerebroventricular injection in either the tail-pinch test or the hot plate test, regardless of the drug tested (data not shown).

Caffeine at a concentration of 1 nmol inhibited the increase in latencies induced by oxymorphone (tail-pinch test, −40%, P < 0. 01; hot plate test, −31%, P < 0.05) and by SNC80 (tail-pinch test, −28%, P < 0.05; hot plate test, −80%, P < 0. 01), whereas caffeine alone (1 nmol) had no effect (fig. 2).

Adenosine uptake was evaluated by measuring the decrease in adenosine concentration in the supernatant as a function of time (fig. 3). None of the opioids tested had a significant effect on adenosine uptake at a concentration of 10 nm. Oxymorphone at 100 nm increased adenosine concentration by a mean of 17% compared with saline (P < 0. 01), but at 1,000 nm, it increased adenosine concentration by a mean of 15% compared with 100 nm (fig. 3A;P < 0.05). SNC80 at 100 nm increased adenosine concentration by a mean of 30% compared with saline, but at 1,000 nm, it increased adenosine concentration by 19% compared with 100 nm (fig. 3B;P < 0.05). U69593 had no significant effect on adenosine concentration in the supernatant, regardless of the concentration used (fig. 3C).

In this study, we demonstrated that oxymorphone and SNC80 inhibited adenosine uptake by whole brain synaptosomes in a dose-dependent manner, whereas U69593 did not. This inhibition may increase extracellular adenosine concentration, but the molecular mechanism by which opiates interact with adenosine transport remains unknown. We also found that injecting 1 nmol caffeine was sufficient to inhibit partially the increase in latency in the tail-pinch and hot plate tests induced by oxymorphone or SNC80, whereas 1 nmol caffeine alone had no effect (at higher concentrations, caffeine may have analgesic effects). 16U69593 had a weaker analgesic effect than oxymorphone or SNC80 and did not inhibit adenosine uptake in brain synaptosomes. The weaker analgesic effect of this κ agonist was presumably due to its inability to inhibit adenosine uptake and to increase adenosine concentration in the extracellular spaces. Consistent with this notion is our observation that caffeine did not affect the latency increase induced by the κ agonist in the hot plate and tail-pinch tests.

The multiple mechanisms of action of caffeine may explain the complex and controversial interactions between caffeine and opiates. Caffeine in rats potentiates the morphine analgesia in a dose-dependent manner, while in mice, low-dose caffeine inhibits the analgesic effects of morphine (Malec and Michalska, 17our study). Also, high- but not low-dose caffeine potentiates morphine-induced analgesia, 18and caffeine alone has analgesic effects via  the amplification of cholinergic transmission. 19Caffeine acts via  many mechanisms, including the inhibition of phosphodiesterases, 20but most pharmacological effects result from antagonism of adenosine receptors, caffeine acting most potently at A2A, followed by A1 and then A2B receptors. 21At a high dose, caffeine can modulate many other receptor systems.

Inhibition of the neurotransmitter release induced by morphine is mediated by the initial release of adenosine. 3Also, oxymorphone and SNC80 inhibit adenosine uptake and increase adenosine concentration in rat striatum, whereas U69593 does not. 13In this study, we found that oxymorphone and SNC80 inhibited adenosine uptake by whole brain synaptosomes, whereas U69593 did not. Thus, overall, these results suggest that the activation of a μ or δ receptor leads to an increase in extracellular adenosine concentration and that this may be a general mechanism of interaction at supraspinal sites because adenosine is taken up via  a facilitated diffusion system that is present in most mammalian cells. 4 

Uptake inhibition is probably not the only mechanism involved in the increase in adenosine concentration in the extracellular spaces that is induced by opioid receptor agonists. Mu- and to a lesser extent δ-opioid agonists increase adenosine release via  a Ca²⁺-dependent pathway, whereas κ agonists have little effect on adenosine release. 10Our results are consistent with these findings. Also, adenosine release is mediated by the activation of N-type voltage-sensitive Ca²⁺channels. 22,23However, in our study, it was not possible to differentiate between the effects of opioids on Ca²⁺-dependent adenosine release and uptake.

Finally, we found that the A1 adenosine receptor agonist was more efficient than the A2 receptor agonist, in terms of antinociceptive effects, following intracerebroventricular injection. This ranking order has been reported for the spinal level in previous studies. 24Delander et al. , 25using NECA as an A2 agonist, showed that the A2 receptor agonist had a greater effect than the A1 receptor agonist. However, NECA is a nonselective agonist 26and therefore cannot be used to implicate adenosine A2 receptors specifically in analgesia.

In summary, we found that oxymorphone and SNC80 (μ and δ agonist, respectively) inhibited the uptake of adenosine by brain synaptosomes, whereas U69596 (a κ agonist) did not. Furthermore, caffeine, a nonspecific adenosine receptor antagonist, decreased the antinociceptive effects obtained with oxymorphone or SNC80 but had no effect on those obtained with U69593 in acute pain models. Our results suggest that the analgesic effects of μ and δ agonists involve an increased adenosine availability, possibly due to the modulation of adenosine uptake.

Two points limit some of our work. First, even if facilitated diffusion systems appear to be a general uptake mechanism in mammalian cells, it would be interesting to evaluate the effect of opioid agonists on adenosine uptake in brain areas implicated in the control of pain.

Second, even if the principal result here concerns the modulation of adenosine uptake by specific opiate agonists, the use of specific adenosine receptor agonists in vivo  might bring information on the subtype of purinergic receptor interacting with the opioid.