Clonidine produces analgesia after spinal injection by activating alpha2-adrenergic receptors. Recently, clonidine has been demonstrated to increase spinal release of norepinephrine (NE) in vivo, in contrast to that anticipated by classic presynaptic autoinhibition. The purpose of the current study was to determine if clonidine could inhibit release of NE in a preparation of spinal cord tissue lacking synaptic circuits.
Crude synaptosomes were prepared from male Sprague-Dawley rat spinal cord, loaded with [3H]NE, and stimulated by potassium chloride to release [3H]NE. Samples were incubated with clonidine in the absence or presence of various inhibitors. To study the effect of alpha2a-adrenergic receptor subtypes, some animals were pretreated with an oligodeoxynucleotide (ODN) composed of a sense or antisense sequence to a portion of this receptor.
Potassium chloride produced a concentration-dependent increase in [3H]NE release, and this release was inhibited by clonidine with a concentration producing 50% maximal inhibition (IC50) of 1.3 microm. The effect of clonidine was inhibited by the alpha2-adrenergic antagonists, yohimbine and idazoxan, but not by alpha1-adrenergic, muscarinic, or opioid antagonists. Intrathecal pretreatment with antisense ODN to alpha2A-adrenergic receptors reduced alpha2A-adrenergic receptor protein expression compared with sense ODN control and also reduced clonidine-induced inhibition of [3H]NE release.
These data demonstrate the existence of classic autoinhibitory alpha2-adrenergic receptors in the spinal cord, probably of the alpha2Asubtype. They further suggest that clonidine-induced stimulation of spinal NE release must occur from indirect actions, presumably due to activation of a spinal circuit.
INTRASPINAL administration of α2-adrenergic agonists produces antinociception in animals, 1and intraspinal administration of the α2-adrenergic agonist, clonidine, is effective in a variety of acute and chronic pain states in humans. 2These agents produce antinociception by stimulating α2-adrenergic receptors, 3by inhibiting release of excitatory amino acids and substance P from central terminals of primary afferents, 4,5and by hyperpolarizing spinal cord neurons. 6However, whether these are direct actions at these targets or indirect via activation of inhibitory circuits or inhibition of excitatory circuits is not addressed by these studies. Gordh et al. 7first suggested that α2-adrenergic agonists produce antinociception by stimulating spinal cholinergic neurons. Further experiments suggest the existence of a circuit by which α2-adrenergic receptor stimulation results in spinal cholinergic interneuron stimulation and activation of nitric oxide synthase in producing antinociception. 8
Norepinephrine (NE) is the neurotransmitter that stimulates α2-adrenergic receptors in the spinal cord. α2-Adrenergic receptors are located on noradrenergic terminals in the peripheral and central nervous system and act in an autoinhibitory manner to diminish further NE release. 9,10In contrast, intrathecal administration of α2-adrenergic agonists increases NE in cerebrospinal fluid of humans and in spinal cord microdialysates of animals. 11,12This would suggest a lack of autoinhibitory α2-adrenergic receptors in the spinal cord. Immunocytochemical staining for α2-adrenergic receptors does not colocalize with noradrenergic fibers in the spinal cord, 13consistent with such a lack. Similarly, a previous study in spinal cord slices provided presumptive evidence for α2-adrenergic autoinhibitory receptors but revealed clonidine-induced inhibition of NE release only at high clonidine concentrations (0.01–2.5 m), 14which could reflect nonspecific actions. The purpose of the current study was to determine, using a synaptosomal preparation lacking intact neuronal circuits, whether lesser concentrations of clonidine affected NE release and by what receptor mechanism clonidine produces such an effect.
Materials and Methods
After obtaining approval from the Animal Care and Use Committee, male Sprague-Dawley rats (250 g) were studied. After induction of anesthesia with 1.5–2.1% inhalational halothane, the animals were killed by decapitation, and the spinal cord was quickly removed and placed in aerated (with 95% O2/5% CO2) ice-cold modified Krebs-bicarbonate buffer containing 118 mm NaCl, 3.3 mm KCl, 1.2 mm MgSO4, 1.25 mm CaCl2, 1.2 mm KH2PO4, 25 mm NaHCO3, 10 mm HEPES, 5 mm ascorbic acid, 11.5 mm glucose, 30 μm EDTA, and 10 μm pargyline. The dorsal half of the spinal cord was selected and homogenized in 8 ml of ice-cold 0.32 m sucrose. A crude synaptosomal pellet (P2) was prepared by differential centrifugation at 2,000 g followed by 20,000 g . 15
The crude P2pellet was resuspended into 4 ml of modified Krebs buffer, loaded with NE in a 50 nm final concentration containing 20%[3H]NE, and incubated at 37°C for 5 min. The free NE was then removed by centrifugation at 15,000 g for 10 min. The synaptosomal pellet was again suspended into 4.5 ml of modified Krebs buffer, and 150 μl of the suspension was aliquoted into each test tube with 850 μl Krebs buffer containing altered KCl concentration alone or with clonidine, various antagonists, or their combinations. The test tubes were then incubated for 10 min at 37°C in a 1-ml volume. At the end of incubation, the amount of [3H] remaining in synaptosomes was determined by rapid filtration through GF/C Glass fibers (Corning, Corning, NY) presoaked for 30 min or more in 0.1% (vol/vol) polyethylenimine to reduce nonspecific binding. This was followed by three times 4-ml washes with ice-cold buffer in which glucose was substituted for NaCl. The bound (retained) radioactivity was determined 24 h later by 1219 Rack Beta Scintillation Counter (LKB, Wallac Inc, Gaitherburg, MD) in Bio Safe II scintillation fluid. [3H]NE release induced by KCl was calculated from the amount of [3H]NE remaining in the synaptosome after vehicle (100 μl buffer) compared with altered KCl treatment. Preliminary experiments confirmed the work of others regarding the incubation timing and temperature of these experiments. 15
Animals were anesthetized, and a polyethylene catheter was inserted through a nick in the cisternal membrane and advanced 8.5 cm such that its tip lay in the lumbar intrathecal space as previously described. 16Rats showing neurologic deficits after awakening were euthanized immediately. To confirm correct placement of the catheters, 10 μl of 2% lidocaine was injected, followed by a 10-μl saline flush the day after surgery. Only animals that developed transient bilateral motor and sensory blockade in the hind legs were included in the study. After implantation of the intrathecal catheters, rats were housed individually with free access to food and water and allowed to recover for 5 days before use. Animals then received intrathecal injection of an oligodeoxynucleotide (ODN) representing an antisense or sense composition specific for the α2A/D receptor, for which the antisense has been demonstrated to reduce receptor expression when administered in brain. 17ODN treatment consisted of twice-daily injections for 3 days in a dose of 5 nmol per injection (5 μl volume, followed by 10 μl saline to flush catheter deadspace). This treatment schedule results in loss of antinociception for intrathecally administered clonidine (X Paqueron, manuscript in preparation). Animals were deeply anesthetized with phenobarbital and decapitated, and spinal cords were rapidly removed and frozen in 2-methylbutane in the presence of dry ice and stored at −70°C.
Western Blot Analysis
Spinal cords were divided into dorsal and ventral halves, and the dorsal half was used for analysis. The tissue was chopped on dry ice, homogenized in ice-cold lysis buffer (10 mm Tris, pH 7.5, 1% sodium dodecylsulfate), denatured at 100°C for 5 min, chilled on ice, and centrifuged. The protein concentration in the supernatant was determined (DC protein assay; Bio-Rad, Hercules, CA), and samples were diluted to make protein concentrations equal. Samples (50 μg protein) were then loaded on a sodium dodecylsulfate–polyacrylamide gel and electrophoresed at 110–170 V in running buffer (0.025 m Tris, 0.25 m glycine, 0.1% sodium dodecylsulfate). Proteins were transferred to a nitrocellular membrane using a transfer unit. Membranes were washed three times (5 min each) with phosphate-buffered saline containing 0.05% Tween 20, then incubated for 60 min with this buffer plus 5% nonfat dry milk. Membranes were then incubated with an antibody to the α2A/D-adrenergic receptor (Research Biochemicals, Inc, Natick, MA) in a 1:5,000 dilution at room temperature for 1 h, washed with phosphate-buffered saline–Tween buffer, then incubated with a peroxidase-conjugated secondary antibody (Jackson ImmunoResearch Lab, Inc., West Grove, PA). Protein bands were visualized by means of an ECL detection kit (Amersham, Piscataway, NJ). The specificity of this antibody, raised against amino acids 218–235 of the α2A/D receptor, an area unique to this subtype, was confirmed by Western immunoblotting of protein purified from fibroblasts expressing only the α2A/D- or only the α2C-adrenergic receptor. Densitometric analysis was quantified using NIH Imaging software (National Institutes of Health, Bethesda, MD).
To compare results across gels, all densitometric readings on a given gel were compared with a 50-μg protein sample on the same gel that was prepared from homogenates of fibroblasts transfected to express the α2D-adrenergic receptor. These cells, as well as those expressing the α2C-adrenergic receptor, were a generous gift from S. M. Lanier, who derived and characterized them. 18Protein was prepared from these cells at one time and frozen at −70°C in 50-μg aliquots, one of which was thawed for each gel that was performed.
l-[2,5,6-3H]NE (62 Ci/mmol) was purchased from New England Nuclear (Wilmington, DE). Bio Safe II scintillation cocktail was obtained from Research Product International Corp (Mount Prospect, IL). MgSO4, ascorbic acid, KCl, and glucose were obtained from Fisher Scientific (Fair Lawn, NJ). The remaining chemicals were obtained from Sigma Chemical Co. (St. Louis, MO).
All release experiments were performed in four sets in duplicate, and values are expressed as mean ± SEM. The percentage release of NE was calculated by dividing the loss of radioactivity in each sample against the basal radioactivity in the control sample without KCl. Densitometric data on each gel were expressed as percent of optical density from the 50-μg protein sample loaded on that gel from α2D-adrenergic receptor expressing transfected fibroblasts. Data were analyzed by one- or two- way analysis of variance, with P < 0.05 considered significant.
Addition of KCl produced a concentration-dependent increase in [3H]NE release, with 3.9 ± 0.9%, 16.3 ± 0.8%, 29.7 ± 1.2%, and 36 ± 2.5% release after 5, 10, 30, and 100 mm KCl, respectively (P < 0.05 vs control in each case). Clonidine (1 μm) reduced [3H]NE release from the lower KCl concentrations (fractional release of 0.4 ± 2.1%, 13 ± 11%, and 25 ± 1.9% after 5, 10, and 30 mm KCl, respectively, P < 0.05 vs KCl without clonidine). However, [3H]NE release was not affected by this concentration of clonidine after exposure to the highest concentration of KCl, 100 mm (fractional release of 35 ± 2.5%; not significant vs KCl without clonidine). In subsequent studies, [3H]NE release was studied from 10 mm KCl. Using this degree of stimulation, clonidine produced a concentration-dependent inhibition of [3H]NE release with a concentration producing a 50% maximal effect (IC50) of 1.3 ± 2.6 μm (fig. 1).
A series of studies were then performed using a stimulating concentration of KCl for [3H]NE release of 10 mm and clonidine, 1 μm to investigate the receptors by which clonidine produced its inhibition. Incubation with yohimbine alone, 1 nm to 10 μm, had no effect on KCl-induced [3H]NE release (data not shown), but the highest yohimbine concentration completely blocked clonidine’s inhibition of [3H]NE release (fig. 2), consistent with an action on α2-adrenergic receptors. When the more potent α2-adrenergic antagonist idazoxan was examined in this same concentration range (1 nm to 10 μm), concentrations greater than 100 nm stimulated [3H]NE release above that from KCl alone (10 μm idazoxan increased release 19 ± 1.9% above that from KCl alone). However, 100 nm idazoxan, which did not affect [3H]NE release alone, completely blocked inhibition from clonidine (fig. 3), again consistent with an action on α2-adrenergic receptors.
Neither atropine (muscarinic antagonist) nor naloxone (opioid antagonist) altered KCl-induced [3H]NE release alone (data not shown) or affected clonidine’s inhibition of [3H]NE release (fig. 4). Corynanthine, a specific α1-adrenergic antagonist, in concentrations up to 10 μm, had no effect on [3H]NE alone and did not affect clonidine-induced inhibition of [3H]NE release (fig. 5, top). Unexpectedly, prazosin alone at concentrations greater than 10 nm further stimulated [3H]NE release above that from KCl alone (fig. 5, bottom). Prazosin concentrations up to 10 nm did not affect clonidine-induced inhibition of [3H]NE release (fig. 5, bottom).
We next examined the α2-adrenergic receptor subtypes responsible for the effect of clonidine. First, ARC 239, an antagonist relatively selective for α2B/C-adrenergic receptors, 19,20was studied. This agent, in concentrations greater than 100 nm, stimulated [3H]NE release (51 ± 2.6% release at 10 μm). However, in concentrations that did not affect [3H]NE release, ARC 239 had no effect on clonidine’s inhibition of [3H]NE release (fig. 6). Second, we examined the effect of α2A-adrenergic receptors by the intrathecal antisense ODN exposure. Treatment with antisense reduced α2A-adrenergic receptor protein by 61 ± 12% compared with sense treatment, as determined by Western analysis (P < 0.05; example of immunoblot in fig. 7, top). The clonidine concentration-dependent inhibition curve was shifted to the right by more than two orders of magnitude in antisense- compared with sense-treated animals (fig. 7, bottom;P < 0.001 by two-way analysis of variance), indicating an action on α2A-adrenergic receptors.
The current study is consistent with a series of previous investigations that used other methods to determine the presence of presynaptic α2-adrenergic receptors in the spinal cord but provides unique information on the role of such receptors regulating local NE release. That spinal cord contains presynaptic α2-adrenergic receptors on the central terminals of primary afferent fibers is supported by a reduction in α2-adrenergic radioligand binding in the dorsal horn after rhizotomy, 21a reduction in stimulated glutamate, substance P, and calcitonin gene-related peptide release by incubation with α2-adrenergic agonists in vitro , 4,5,22and a reduction in stimulated glutamate release in synaptosomes by α2-adrenergic agonists. 23
Evidence for α2-adrenergic receptors on noradrenergic terminals in the spinal cord is less compelling than that previously cited for primary afferents. There is a minor reduction in α2-adrenergic radioligand binding in spinal cord after spinal cord transsection (which disrupts all noradrenergic innervation of the spinal cord), 24consistent with the existence of a low level of presynaptic α2-adrenergic receptors. Co-immunostaining of spinal cord slices with antibodies specific to α2-adrenergic receptors (subtypes A or C) and antibodies specific to noradrenergic fibers (dopamine β hydroxylase) fails to demonstrate such antibodies on noradrenergic fibers by confocal microscopy. 13However, the sensitivity of the immunocytochemical technique to detect a low level of expression of α2-adrenergic receptors is uncertain, and this result does not exclude the presence of such presynaptic α2-adrenergic receptors.
Other studies are similarly equivocal concerning the presence of autoinhibitory α2-adrenergic receptors in the spinal cord. In a previous study of spinal cord slices, clonidine inhibited KCl-induced NE release with a pharmacology consistent with an action on α2A-adrenergic receptors, as it was not inhibited by ARC 239. 14However, the clonidine concentrations used in that study (0.01–2.5 m) are considerably greater than those used to demonstrate the existence of autoinhibitory α2-adrenergic receptors at supraspinal sites, 25raising concerns of possible nonspecific effects. The reason for the discrepancy between the study by Umeda et al. 14and are own are unclear, but the current study demonstrates clonidine inhibition at nearly 1,000-fold lesser concentrations.
Clonidine was less effective at inhibiting NE release to higher than lower KCl concentrations in the current study. This is consistent with a previous report in postganglionic sympathetic noradrenergic neurons in culture, where autoinhibition by α2-adrenergic stimulation is less effective (percent reduction is decreased and effective concentration producing 50% inhibition is increased to α2-adrenergic agonists) as electrical stimulation intensity increases. 9Reduction in presynaptic autoinhibition with increasing stimulation intensity has been a general finding and has been ascribed to increasing levels of intraaxonal Ca2+. 26
Clonidine could inhibit NE release by direct actions on noradrenergic synaptosomes or by stimulating other synaptosomes to release inhibitory neurotransmitters, which could bind to heterotopic presynaptic receptors on noradrenergic synaptosomes. 27Two such heterotopic receptors are muscarinic and opioid, yet we found no evidence for such a mechanism, since neither atropine nor naloxone affected clonidine-induced inhibition. Although other interpretations are possible, we believe that blockade of clonidine’s effect by yohimbine and idazoxan, and increase in NE release by high concentrations of the more potent α2-adrenergic antagonists, idazoxan and ARC 239, alone most likely reflect the presence of presynaptic autoinhibitory α2A-adrenergic receptors. Concentrations for inhibition by idazoxan and yohimbine of clonidine’s effect are consistent with those previously observed. 28We did not include in these experiments inhibitors of NE reuptake because α2-adrenergic receptors do not directly interact with them; therefore, we cannot exclude the unlikely possibility of a dual action on NE release and reuptake.
The lack of effect of corynanthine or low concentrations of prazosin on the effect of clonidine suggests this is not due to an action of clonidine on α1-adrenergic receptors. Stimulation of NE release by prazosin is unexpected and is most likely a nonspecific effect, because it was not mimicked by corynanthine. It is unlikely that α1-adrenergic receptors would inhibit NE release because there is no precedence for autoinhibitory α1-adrenergic receptors in the central or peripheral nervous system. There are α1-adrenergic receptors in the spinal cord, 29and activation of such receptors can stimulate cholinergic motor neurons 30and preganglionic sympathetic neurons, 31as well as facilitate responses to noxious stimuli. 32
Both α2Aand α2Csubtypes of adrenergic receptors exist in rat spinal cord. 13Most studies implicate the αA subtype in analgesia, 33,34although this is not universal. 35Interpretation of studies in mice lacking the gene for receptor subtypes, as performed in many of these examples, is complicated by the unknown effects of lack of a central nervous system receptor throughout development. For this reason, many investigators prefer the use of antisense ODN treatment to acutely knockdown receptor protein expression. A previous such study in rats suggested the α2A subtype was responsible for antinociception, 36although the investigators in that study did not demonstrate acute knockdown of α2A-adrenergic receptor protein in the spinal cords of those animals by their treatment. The current study did not address the α2-adrenergic receptor subtype responsible for antinociception in the rat. However, using this antisense ODN approach, the current study does demonstrate acute knockdown of α2A-adrenergic receptor protein in spinal cords of rats and simultaneous reduction in the potency of clonidine to inhibit NE release, suggesting that at least some of the autoinhibitory α2-adrenergic receptors are of the α2A subtype. These results are consistent with pharmacologic studies in mouse postganglionic sympathetic neurons (and review in this reference of rabbit, mouse, and guinea pig atria and cerebral cortex) 9that autoinhibitory α2-adrenergic receptors are of the A/D subtype. The potency of ARC 239 at α2B/C receptors is at least 100 times more than at α2A receptors, suggesting that lack of inhibition at the concentrations used in this study (as also observed in locus coeruleus) 37reflects either no autoinhibitory α2-adrenergic receptors of the α2B/C type, or a very small component of them.
As noted in the Introduction, intrathecal administration of α2-adrenergic agonists does not decrease NE in circulating cerebrospinal fluid or in microdialysates from sheep as predicted by the presence of autoinhibitory α2-adrenergic receptors, but increases such concentrations. 11,12We have proposed, based on these studies, a feed-forward circuit of NE →α2-adrenergic receptors → acetylcholine release → muscarinic/nicotinic receptors → nitric oxide synthesis → NE release. It would appear that under normal circumstances this circuit swamps the effect of local autoinhibitory α2-adrenergic receptors. An alternative explanation would be a direct α2-adrenergic agonist–induced release of NE from noradrenergic terminals in sheep, but not rats, which we consider unlikely. The pharmacologic and physiologic stimuli for activation of this feed-forward circuit and the brakes to this circuit are under investigation.
In summary, clonidine reduces KCl-induced NE release from isolated spinal cord synaptosomes with an IC50 of 1.3 μm and a pharmacology consistent with an action on α2-adrenergic receptors. Antisense and pharmacologic studies are consistent, with a major component of clonidine’s inhibition of NE release being due to actions on α2A-adrenergic receptors. Because intrathecal clonidine increases, not decreases, NE release in vivo , the regulation of NE release by α2-adrenergic receptors in the spinal cord is complex and involves both direct and indirect actions on noradrenergic nerve terminals.