α²-Adrenoceptor Activation by Clonidine Enhances Stimulation-evoked Acetylcholine Release from Spinal Cord Tissue after Nerve Ligation in Rats

After approval was obtained from the institutional Animal Care and Use Committee (Wake Forest University School of Medicine, Winston-Salem, North Carolina), male Sprague-Dawley rats weighing 200–280 g at the time of purchase (Harlan, Indianapolis, IN) were studied. Animals were housed under a 12-h light–dark cycle, with access to food and water ad libitum . Animals were separated into two groups: no surgery (n = 35) or SNL (n = 66). We chose no surgery as a control rather than sham surgery with exposure of nerves without ligation, because we were interested in changes within the dorsal root ganglia cells during the neuropathic state. We recognize that we therefore could not exclude an effect of inflammation or surgery itself as opposed to nerve injury.

Spinal nerve–ligated rats weighing 200–220 g were anesthetized with halothane (2–3% in oxygen), and the left L5 and L6 spinal nerves were isolated and ligated tightly with 6-0 silk sutures distal to the dorsal root ganglia as previously described.12After surgery, animals were housed individually with free access to food and water and were allowed to recover for at least 14 days. Left paw tactile allodynia was confirmed at this time by measuring the hind paw withdrawal threshold in response to application of von Frey filaments, using a previously described up–down method.13Only animals with a withdrawal threshold less than 4 g were used. All in vitro  studies were performed 3 weeks after SNL surgery, at a time of allodynia to von Frey filament testing.

After induction of anesthesia with 2–3% halothane, animals were killed by decapitation, and the lumbar enlargement (approximately L2–L5) of the spinal cord was quickly removed and placed in oxygenated (with 95% oxygen–5% carbon dioxide) ice-cold modified Krebs buffer containing 125 mm NaCl, 3 mm KCl, 1.2 mm MgSO⁴, 1.2 mm CaCl², 1 mm KH²PO⁴, 22 mm NaHCO³, and 11.5 mm glucose (pH = 7.35). The left dorsal quadrant of the spinal cord ipsilateral to nerve injury was removed from SNL rats and homogenized in 8 ml ice-cold 0.32 m sucrose. To decrease the number of rats, the entire dorsal half of the spinal cord was used in normal rats. Each synaptosome preparation contained tissue of four SNL rats or two normal rats. A crude synaptosomal pellet was prepared from mechanically homogenized tissue by differential centrifugation at 2,000g  followed by 20,000g  as previously described.14 

Acetylcholine release from synaptosomes and tissue slices was done after loading with [³H]choline according to standard methods, which results in rapid uptake of choline, acetylation to acetylcholine, and release of [³H]acetylcholine with depolarization.15,16We did not, however, confirm in the current experiments that released radioactivity was completely in the form of [³H]acetylcholine. For the synaptosome experiments, the crude pellet was resuspended into 4 ml modified Krebs buffer, loaded with [³H]choline, and incubated at 37°C for 20 min. Free [³H]choline was then removed by centrifugation at 20,000g  for 5 min. The synaptosomal pellet was again suspended into 4.5 ml modified Krebs buffer, and 150 μl of the suspension was aliquoted into each test tube with 850 μl modified Krebs buffer containing different concentrations of clonidine. Each tube also contained hemicholium-3 at a final concentration of 10 μm to prevent reuptake of acetylcholine. For KCl stimulation, the concentration of KCl in the test tube was increased to12 mm. Tubes were then incubated for 10 min at 37°C in a total volume of 1 ml. At the end of incubation, the amount of [³H] remaining in synaptosomes was determined by rapid filtration through GF/C glass filters presoaked for 30 min in 0.1% polyethylene 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 scintillation counting. The effects of clonidine at doses from 10⁻⁸to 10⁻⁴m on [³H]acetylcholine release were determined by the amount of tritium remaining in the synaptosomes incubated in buffer without high KCl concentration (control) compared with treatment with each concentration of clonidine with physiologic (3 mm) or depolarizing (12 mm) concentrations of KCl. Fractional release was calculated in each sample from the retained radioactivity, and percentage change in fractional acetylcholine release by clonidine defined as (control − clonidine)/control × 100. Each experiment was performed in duplicate for each concentration, and the mean value was included.

To determine whether another α²-adrenoceptor agonist would mimic the effects of clonidine, synaptosomes from normal and SNL rats (n = 7 each) were prepared, and the effect of KCl alone and in the presence of 10 nm dexmedetomidine was examined.

Rats were anesthetized with halothane and killed by decapitation, and the lumbar enlargement of the spinal cord was quickly removed. The left dorsal semihalf was chopped in 0.6-mm slices. The slices were preincubated with [³H]choline for 40 min at 37°C in oxygenated (with 95% oxygen–5% carbon dioxide) modified Krebs buffer containing 125 mm NaCl, 3 mm KCl, 1.2 mm MgSO⁴, 1.2 mm CaCl², 1 mm KH²PO⁴, 22 mm NaHCO³, and 11.5 mm glucose (pH = 7.35). Four to five spinal cord slices were put into each chamber surrounded by a temperature-controlled water bath maintained at 33°C.

The slices were superfused continuously with a multichannel pump at 0.5 ml/min with oxygenated modified Krebs buffer containing 10 μm hemicholium-3. After an initial perfusion of 40 min, 5-min fractions were collected. At 15 min and again at 50 min after beginning of fraction collection, the slices were depolarized for 5 min by increasing the concentration of KCl in the superfusion fluid to 25 mm. The total tritium overflow in response to 25 mm KCl was usually between 150 and 200% of the baseline release.

The test drugs (clonidine and idazoxan) were applied beginning 40 min after the beginning of fraction collection until 55 min. This coincided to the time beginning 10 min before the second KCl depolarization to the end of this depolarization. The radioactivity of each sample was determined 24 h later by scintillation counting. The KCl-evoked [³H]acetylcholine release was calculated by subtracting the values of the basal release from the total release during the stimulation period. For each sample, the ratio of fractional release value during the second to the first KCl depolarization was calculated. In control slices, this ratio was 1.01 ± 0.05. The effects of clonidine concentrations from 10⁻⁶to 10⁻⁴m on KCl-evoked [³H]acetylcholine release were determined and compared to release in the presence of buffer only before KCl depolarization.

[³H]Choline was purchased from Perkin Elmer (Boston, MA). Bio Safe II scintillation cocktail was obtained from Research Product International Corp. (Mount Prospect, IL). MgSO⁴, KH²PO⁴, and NaHCO³were obtained from Fisher Scientific (Fair Lawn, NJ). The remaining chemicals were obtained from Sigma Chemical Co. (St. Louis, MO).

Experiments were replicated four to five times for synaptosomes and five times for slices. All data are expressed as mean ± SEM. For concentration–response studies, groups were compared by one-way or two-way analysis of variance. For antagonist studies, the clonidine-treated group and other groups were compared by one-way analysis of variance. A P  value less than 0.05 was considered significant.

In normal rats, 10⁻⁸to 10⁻⁴m clonidine alone (before KCl depolarization) increased [³H]acetylcholine release compared with vehicle. In contrast, the highest concentration of clonidine inhibited [³H]acetylcholine release in SNL rats compared with vehicle in the absence of KCl depolarization (fig. 1). The effect of clonidine alone on [³H]acetylcholine release differed significantly in synaptosomes between normal and SNL animals.

KCl, 12 mm, provoked [³H]acetylcholine release by 17 ± 0.9% in synaptosomes from normal rats and 17 ± 1.2% in synaptosomes from SNL rats, compared with Krebs buffer control. The effect of clonidine on KCl-evoked [³H]acetylcholine release in synaptosomes from SNL rats was greater than that in synaptosomes in normal rats (fig. 1). In a separate analysis within normal and SNL animals, 1 μm clonidine increased KCl-evoked acetylcholine release compared with control in synaptosomes from SNL animals, whereas 100 μm clonidine inhibited KCl-evoked acetylcholine release compared with control in synaptosomes from normal animals (fig. 1).

Qualitatively similar results were obtained with the highly selective α²-adrenoceptor agonist dexmedetomidine. In spinal cord synaptosomes from normal animals, 10 nm dexmedetomidine inhibited KCl-evoked [³H]acetylcholine release, whereas in spinal cord synaptosomes from SNL animals, dexmedetomidine enhanced KCl-evoked [³H]acetylcholine (fig. 2; P < 0.005 for each effect).

Perfusion of spinal cord slices for 5 min with 25 mm KCl provoked [³H]acetylcholine release at 164 ± 3% of 3 mm KCl control. The second KCl exposure resulted in less fractional release than the first, with a ratio of 0.65 ± 0.019 (n = 45). Compared with vehicle control within each group, clonidine did not alter the response to the second KCl exposure, although there were strong trends (P < 0.1 for main analysis of variance) toward enhanced release from slices from SNL animals and reduced release in slices from normal animals. Comparing the effect of clonidine between the groups demonstrated significantly greater KCl-evoked acetylcholine release in slices from SNL animals than in slices from normal animals, with a peak effect at 10⁻⁵m (fig. 3).

Simultaneous perfusion of 10⁻⁵m idazoxan with 10⁻⁵m clonidine in spinal cord slices from SNL animals significantly reduced KCl-evoked release. Interestingly, idazoxan alone produced a small reduction in response to the second depolarization with KCl (fig. 4), consistent with tonic α²-adrenoceptor tone in this preparation.

Spinally released acetylcholine produces analgesia primarily by actions on muscarinic cholinergic receptors.17Activation of muscarinic receptors facilitates γ-aminobutyric acid release onto primary afferents in lamina II to inhibit the release of glutamate.18,19In addition, evoked activity of projecting neurons in deeper laminae of the spinal cord is inhibited in vivo  by muscarinic receptors.20Spinal acetylcholine release is increased by analgesics, including systemically administered morphine21and intrathecally administered carbachol22and clonidine.23 

The purpose of the current study was to better understand the mechanisms that underlie the increase in potency and efficacy of α²-adrenoceptor agonists for analgesia in chronic compared with acute pain states. The focus on acetylcholine follows from observations that intrathecal clonidine increases concentrations of this neurotransmitter in patients with pain4and that intrathecal clonidine-induced relief of hypersensitivity to mechanical stimuli in animals with peripheral nerve injury is blocked by intrathecal atropine, whereas atropine does not block the effect of clonidine in normal animals.2The selective reliance of clonidine on muscarinic cholinergic receptor activation in animal models of neuropathic pain could reflect increased clonidine-induced acetylcholine release. Alternatively, clonidine induced acetylcholine release could be similar in normal and SNL animals, but there may be a change in cholinergic receptor expression or location. The current study provides evidence that clonidine stimulates spinal acetylcholine release after peripheral nerve injury.

All previous studies examining acetylcholine release by α²-adrenoceptor agonists were performed in vivo , by either dorsal horn microdialysis or sampling of cerebrospinal fluid. The current study sought to determine whether clonidine-induced acetylcholine release reflected an action on nerve terminals themselves, which we tested using synaptosomes, or acted on local spinal circuits, which we tested using spinal cord slices. Synaptosomes are commonly used to examine regulation of neurotransmitter release in the absence of indirect effects from disinhibition or activation of excitatory neuronal inputs to the terminals.9,10,24 

Clonidine alone produced a small and concentration-independent release of acetylcholine from spinal cord synaptosomes in normal animals and none from synaptosomes from SNL animals. We did not anticipate that clonidine would induce acetylcholine release in synaptosomes, because α²adrenoceptors are classically inhibitory, coupling with G proteins, which decrease cyclic adenosine monophosphate concentrations and result in reduced voltage-gated calcium channel opening.25Because α²adrenoceptors colocalize with cholinergic neurons in the spinal cord,26we anticipated that clonidine would, if anything, reduce acetylcholine release and that clonidine-induced acetylcholine release in vivo  reflected activation of circuits rather than direct effects on cholinergic terminals.

In contrast to the minimal effect of clonidine alone on acetylcholine release in synaptosomes, clonidine clearly augmented acetylcholine release in synaptosomes depolarized by KCl only in tissue from SNL animals. We chose KCl to nonselectively stimulate all synaptosomes to examine the influence of clonidine on evoked release. As with release from drug alone, we anticipated that clonidine would, if anything, reduce evoked acetylcholine release, not increase it, but this occurred only in tissue from normal animals. The current results, in contrast to our expectations, are consistent nonetheless with reports that α²adrenoceptors can shift from inhibitory to excitatory mechanisms in neurons after peripheral nerve injury27or inflammation,28possibly because of a shift in G-protein species, because α²adrenoceptors can couple with excitatory, Gs species when these are present in adequate numbers.29 

Perfusion of spinal cord slices in vitro  also demonstrated enhancement of evoked acetylcholine release only in tissue from SNL animals. In both synaptosome and slice studies, the effect of clonidine was reduced at higher concentrations, potentially because of actions of clonidine on α¹adrenoceptors.30At a clonidine concentration producing a peak effect, however, this was completely reversed by the selective α²-adrenoceptor antagonist idazoxan. The peak effect of clonidine synaptosomal experiments was an order of magnitude less (10⁻⁶m) than that in slices (10⁻⁵m), possibly reflecting reduced tissue penetration in the slices.

The cause of this plasticity in evoked acetylcholine release from spinal cord after nerve injury is unknown. Previous studies suggest that clonidine inhibits stimulation-evoked neurotransmitter release via  actions on an α^2A-adrenoceptor subtype.31,32However, nerve injury reduces α^2A- but not α^2C-adrenoceptor immunolabeling in the spinal cord, and α^2C-adrenoceptors are located in the deep dorsal horn, the normal termination of large-diameter fibers that subserve mechanical input33and the site of cholinergic cell bodies, which coexpress the α^2Cadrenoceptor.26The α²-adrenoceptor subtype activated by clonidine to reduce mechanical allodynia after nerve injury shifts from the α^2A- to α^2C-adrenoceptor subtype,34and it is possible that increased α^2C-adrenoceptor number or coupling efficiency on spinal cholinergic neurons after nerve injury is responsible for increased evoked acetylcholine release by clonidine.

In summary, clonidine enhances K⁺-evoked acetylcholine release in spinal cord tissue from SNL, but not normal rats, using two different preparations to measure neurotransmitter release: synaptosomes, which only measure direct action on terminals, and spinal cord slices, in which some circuits are preserved. The facilitating effect of clonidine in slices is blocked by the α²-adrenoceptor antagonist idazoxan. These data suggest that, after nerve injury, activation of spinal α²adrenoceptors induces more acetylcholine release from cholinergic neurons when these are stimulated. This neuronal plasticity may contribute to the increased potency of clonidine for analgesia and its dependency on cholinergic neuronal activation in the neuropathic pain state.