Moxonidine, a Selective Imidazoline–α²-Adrenergic Receptor Agonist, Produces Spinal Synergistic Antihyperalgesia with Morphine in Nerve-injured Mice

Experimental subjects were 25–30-g male Institute of Cancer Research mice (Harlan, Madison, WI). Subjects were housed in groups of 5–10 in a temperature- and humidity-controlled environment. Subjects were maintained on a 12-h light–dark cycle and had free access to food and water. Each animal was used only once. These experiments were approved by the Institutional Animal Care and Use Committee.

Moxonidine [4-chloro-5-(2imidazolin-2-ylamino)-6-methoxy-2-methylpyrimidine] chloride was a gift from Solvay Pharma (Hannover, Germany) and was dissolved in 1% acetic acid and diluted with acidified saline (pH 3.2–4). All other drugs were dissolved in 0.9% saline. Morphine was a gift from the National Institute on Drug Abuse (Bethesda, MD). Clonidine HCl (2-[2,6-dichloroaniline]-2-imidazoline) was a gift from Boehringer-Ingelheim Ltd. (Ridgefield, CT). All drugs and drug combinations were injected intrathecally by direct lumbar puncture. 11Briefly, each mouse is gripped firmly by the pelvic girdle. A 30-gauge needle connected to a 50-μl Hamilton syringe is lowered at a 30° angle and inserted at the level of the cauda equina. Puncture of the dura is indicated by a reflexive flick of the tail.

Hypersensitivity was induced by surgical ligation of the L5 spinal nerve in mice. 10Mice were placed in an enclosed chamber and anesthetized by halothane and placed in a prone position before any surgery. When the animal was unresponsive to paw pinch, it was removed from the chamber, shaved from below the iliac crest to approximately halfway to the shoulders, and fitted with a facemask delivering 2 or 3% halothane, which was continuously administered to the animal throughout the surgery. Betadine was applied to the shaved area before the incision. The left paraspinal muscle was separated from the spinous processes at the L4–S2 levels and removed. Removal of this muscle does not impair mobility of the animal after surgery. A Mini-Goldstein retractor (Fine Science Tools No. 17002-02, Foster City, CA) with a 1-cm maximum spread was then inserted into the incision at the level of the iliac crest. Further removal of muscle and tissue permitted visualization of the L6 transverse process and the rostral tip of the sacrum. The L6 transverse process was then removed with use of an S&T fine forceps with a tip dimension of 0.3 × 0.25 mm (Fine Science Tools No. 00108-11). Removal of the process permitted visual identification of the L4–L5 spinal nerves. The L5 spinal nerve was tightly tied (ligated) with 6-0 silk thread distal to the dorsal root and proximal to the confluence of spinal nerves L4 and L5. After hemostasis was confirmed, the wound was sutured with 3-0 silk thread, and the skin was closed with sterile wound clips. The animal was then placed in a moderately heated oxygen-enriched plastic enclosure to facilitate recovery. The animals were fully mobile within 30 min of cessation of anesthetic. As a control, in a separate group of animals, a sham surgery identical to the aforementioned one (but without nerve ligation) was performed.

Nociception was evaluated by responsiveness to multiple applications (10 per hind paw) of a single von Frey filament to the plantar surface of each hind paw. When the stimulus is of sufficient force, the mouse will lick, withdraw, or shake the paw; this action represents the behavioral end point. In nerve-injured mice, a von Frey filament (#3.61) exerting 3.3 mN of force elicited 66 ± 1.3% responsiveness [(number of withdrawals/10)×100] on the paw ipsilateral to the injury. This level of response is sufficient to test compounds for dose-dependent inhibition of the response to mechanical stimulation.

Varying doses of moxonidine, morphine, or clonidine, or combinations thereof, were administered to test for inhibition of tactile sensitivity. Percent inhibition was determined relative to the mean number of paw withdrawals elicited by force and according to the following equation:

% Inhibition = (no. of paw withdrawals predrug − no. of paw withdrawals postdrug)/no. of paw withdrawals predrug

Each mouse served as its own control. The ED⁵⁰values (the dose calculated to produce 50% inhibition) and 95% confidence limits were calculated according to the method of Tallarida and Murray. 12To test for the antihyperalgesic effects of moxonidine and morphine over time, groups of mice injected with various doses of drug or acidified saline were concurrently tested at 5, 10, 30, 60, 90, and 120 min after intrathecal injection. ED⁵⁰values were calculated at the 10-min time point. To test for drug interactions, a separate group of animals (n = 126) was subjected to surgery within the same week. All behavioral testing was conducted the following week on the corresponding day 8 at 10 min after drug injection. New dose–response curves were generated for each drug given alone (morphine, moxonidine, clonidine) or given in combination (morphine–moxonidine, morphine–clonidine), and corresponding ED⁵⁰values were calculated (n = 4–8 mice/dose).

Data describing antihyperalgesia are expressed as means of percent inhibition with SEM. Student t  test comparisons were made between responses of the left and right hind paws of all animals before surgery, and left and right hind paws of nerve-injured, sham, and naive animals after surgery (P < 0.05). The comparison between the left (injured) hind paws of nerve-injured and the left hind paws of sham-operated and naive animals after surgery was also evaluated by analysis of variance. Drug potency comparisons are based on the calculated ED⁵⁰values for the dose–response curve of each drug or combination of drugs.

To test for drug interactions, isobolographic analysis was applied. 12,13When testing an interaction between two drugs given in combination for synergy, additivity, or subadditivity, a theoretical additive ED⁵⁰value is calculated for the combination based on the dose–response curves of each drug administered separately. This theoretical value is then compared by a t  test (P < 0.05) with the observed experimental ED⁵⁰value for the combination. These values are based on total dose of both drugs, i.e. , the dose of clonidine or moxonidine plus the dose of morphine. For the purpose of comparison to the drug doses administered separately, we separated the clonidine or moxonidine and morphine components of the observed and theoretical ED⁵⁰values; these are presented in tables 1 and 2. An interaction is considered synergistic if the observed ED⁵⁰value is significantly less (P < 0.05) than the calculated theoretical additive ED⁵⁰value. 12,13Additivity is indicated when the theoretical and experimental ED⁵⁰values do not differ.

No difference was observed in baseline percent response to a force of 3.3 mN (von Frey filament #3.61, our calibration) between the left (mean = 27 ± 1.8%, n = 142) and right hind paws (mean = 27 ± 1.8%, n = 142;P > 0.05, Student unpaired t  test) of mice before injury. On day 8 after surgery, a substantial increase in responsivity was observed for both hind paws (fig. 1), and the increase was significantly greater for the left hind paw (ipsilateral to the ligation, mean = 66 ± 1.3%, n = 126) than for the right hind paw (contralateral to the ligation: mean = 48 ± 1.8%, n = 126;P < 0.01, Student unpaired t  test;Fig. 1). This small increase in sensitivity on the contralateral side is consistent with previous reports of contralateral effects after nerve injury. 14Both of these responses were substantially greater than that of either hind paw of the control animals; controls included those mice that received sham surgery (left hind paw: mean = 35 ± 15%, n = 6; right hind paw, mean = 33 ± 8.4%, n = 6) and naive mice (left hind paw: mean = 30 ± 6.2%, n = 9; right hind paw: mean = 33 ± 9.9%, n = 9). These differences show that the L5 spinal nerve ligation surgery is sufficient to produce hyperalgesia in the hind paw ipsilateral to the injury.

Moxonidine inhibition of mechanical hyperalgesia is represented in figure 2and expressed as percent inhibition of the percent response to mechanical stimulation. Moxonidine at 0.1- and 1-nmol doses significantly attenuated the hyperalgesia for 10 and 90 min, respectively, whereas 0.03 nmol moxonidine and acidified saline had minimal effect on hyperalgesia. Moxonidine appeared to have a longer duration of action in the ipsilateral hind paw relative to the contralateral hind paw. The calculated ED⁵⁰values of moxonidine at the 10-min time point were comparable between the ipsilateral and contralateral hind paws (ipsilateral: 0.12 nmol, 0.058–0.24; contralateral: 0.12 nmol, 0.037–0.39). We evaluated the doses at the 10-min time point because that time represents the peak analgesic effect at a time most likely involving a selectively spinal effect. 11 

Morphine inhibition of mechanical hyperalgesia is represented in figure 3. Morphine at 3- and 10-nmol doses significantly attenuated the hyperalgesia for the duration of the test period (120 min) in both the ipsilateral and contralateral hind paws. The calculated ED⁵⁰values for morphine at the 10-min time point were comparable between the ipsilateral and contralateral hind paws (ipsilateral: 1.1 nmol, 0.5–2.4; contralateral: 2.4 nmol, 0.88–6.4, not significantly different). Morphine appeared to have comparable duration of action in both the ipsilateral and contralateral hind paws.

Intrathecally administered moxonidine (ED⁵⁰: 14 pmol, 4.1–50) and morphine (ED⁵⁰: 64 pmol, 30–135) both inhibited mechanical hyperalgesia (fig. 4A). Based on these ED⁵⁰values, the moxonidine–morphine equi-effective dose ratio used was 1:4. Combination of moxonidine and morphine at this dose ratio resulted in significant leftward shifts in the dose–response curves (i.e. , increased potency) compared with those of each agonist administered separately (fig 4Aand table 1). The coadministration of moxonidine–morphine combinations in mice resulted in antihyperalgesic dose–response curves with ED⁵⁰values significantly less than the calculated theoretical additive values (fig. 4Band table 1). This result indicates a synergistic interaction.

Intrathecally administered clonidine (ED⁵⁰: 4,600 pmol, 1,800–11,000) and morphine (ED⁵⁰: 64 pmol, 30–135) both inhibited mechanical hyperalgesia (fig. 5A). The morphine–clonidine equi-effective dose ratio used was 1:44. Combination of clonidine and morphine at this dose ratio resulted in significant leftward shifts in the dose–response curves compared with those of each agonist administered separately (fig. 5Aand table 2). The coadministration of clonidine–morphine combinations in mice resulted in antihyperalgesic dose–response curves with ED⁵⁰values significantly less than the calculated theoretical additive values (fig. 5Band table 2). This result confirms a synergistic interaction.

We did not detect obvious motor or sedative side effects with use of these doses of moxonidine, morphine, clonidine, and the combinations; however, we have not conducted systematic evaluation of these effects through use of the rotorod or righting reflex assays.

The current study introduces a new antihyperalgesic agent: the imidazoline–α²-AR agonist moxonidine. The study also shows that both the imidazoline–α²-AR agonists moxonidine and clonidine combined with morphine produce spinal antihyperalgesic synergy in nerve-injured mice.

The ability of α²-AR agonists to produce antihyperalgesia in the mechanical von Frey monofilament stimulation test has been previously observed. 15,16Spinal administration of dexmedetomidine, oxymetazoline, and guanfacine resulted in a dose-dependent reversal of the hyperalgesia induced by L5–L6 spinal nerve ligation in rats. 15,16We have now shown that, like these other α²-AR agonists, moxonidine also dose-dependently decreased hyperalgesic paw withdrawals with a potency comparable to that of morphine and greater than that of clonidine. Morphine remains the standard with which other analgesics are compared, and clonidine is the prototypic analgesic α²-AR agonist. Our comparisons of moxonidine to clonidine and morphine in neuropathic pain in mice suggest that the performance of moxonidine in humans as an analgesic and antihyperalgesic agent may compare favorably with that of morphine and clonidine.

The ability of opioid receptor agonists to inhibit hyperalgesia in nerve-injured animals has also been previously evaluated. Two studies 17,18report that systemically and intracerebroventricularly (but not intrathecally) administered morphine inhibited mechanical hyperalgesia in nerve-injured rats. Additionally, intrathecally administered deltorphin II, a δ opioid receptor–selective agonist, showed decreased antihyperalgesic potency and efficacy in nerve-injured rats. 19Other studies with use of thermal stimulation of the tail as the nociceptive stimulus showed that the intrathecal antinociceptive potency of morphine was decreased approximately twofold 20or fourfold 21in the nerve-injured rats relative to their sham-operated controls. Collectively, these data paralleled the clinical observations that neuropathic pain may be less sensitive to opioids than is nociceptive pain. 22–26 

However, there remains disagreement in the clinical literature over opioid resistance in patients with neuropathic pain. 27,28Some reports have shown success with use of opioids to treat neuropathic pain. 27–30Opioids delivered spinally have been shown to be effective in human patients with neuropathic pain. 31–33Consistent with this clinical experience, at least one study showed that the higher efficacy μ opioid receptor–selective agonist, [D-ala(2),N-MePhe(4),Gly-ol(5)] enkephalin (DAMGO), produced full dose-related antihyperalgesia when given intrathecally to nerve-injured rats. 19Additionally, the intrathecally administered combinations of morphine–deltorphin 19and morphine–clonidine 20produced antihyperalgesia and antinociceptive synergy, respectively, in nerve-injured rats.

Unlike the comparable rat studies, 17,18we observed that intrathecal morphine produces antihyperalgesia in nerve-injured mice at doses comparable to those that are effective in sham-operated and naive controls (data not shown). Furthermore, we observed that morphine synergizes with other antihyperalgesic agents in nerve-injured mice, consistent with other studies showing morphine-coadjuvant synergy (morphine–deltorphin, 19morphine–clonidine 20) in nerve-injured rat. Retention of opioid sensitivity during conditions of neuropathic pain agrees with other clinical reports, 28,34,35that opioids are effective as therapeutic agents for neuropathic pain, albeit with higher dose and/or coadjuvant requirements.

Intrathecal coadministration of morphine with moxonidine produced a synergistic antihyperalgesic effect. The observation of moxonidine–morphine synergy concurs with our previous study that showed antinociceptive synergy between intrathecally coadministered moxonidine and morphine. 8This observation shows that the moxonidine–morphine combination alleviates neuropathic pain responses arising from nerve injury.

Originally, we expected that the morphine–clonidine interaction would not be synergistic in neuropathic mice based on three previous observations: (1) clonidine-mediated spinal analgesia requires the α^2AAR in mice 2; (2) α^2A-AR immunoreactivity decreased in rat spinal cord dorsal horn after nerve injury 36; and (3) clonidine antinociceptive effectiveness decreased in nerve-injured rats. 20However, the current study shows that the clonidine–morphine combination produces antihyperalgesic synergy in nerve-injured mice. Similarly, despite decreases in effectiveness of both drugs when given alone, the clonidine–morphine combination produced antinociceptive synergy in nerve-injured rats 20; these results suggest that, despite decreases in α^2A-AR immunoreactivity in rat dorsal horn after nerve-injury, sufficient receptor numbers remain functional to participate in this interaction with morphine. Recent evidence provides support for this assertion by showing increased α^2A-AR mRNA 37and α^2A-AR immunoreactivity 38in dorsal root ganglia of rats subjected to sciatic nerve transections. These results in dorsal root ganglia together with a previous report 36raise the possibility of altered splicing or trafficking of α^2AAR in the neuropathic state. Alternatively, nerve injury may unmask a latent clonidine effect at upregulated α^2CAR. 36This second possibility is supported by in vitro  studies that indicate that clonidine shows comparable affinity for human α^2A- and α^2C-AR subtypes. 5Regardless, the current data support the use of clonidine as a coadjuvant for morphine for the treatment of neuropathic pain.

In summary, the current results show that both moxonidine and clonidine produce spinal antihyperalgesic synergy with morphine in nerve-injured mice. These results concur with previous evaluations of adrenergic agonists in neuropathic pain 15,16and of morphine–clonidine interactions in normal rodents 39–41and nerve-injured rats. 20This is the first study to show an antihyperalgesic property of the imidazoline–α²-AR agonist moxonidine. It is noteworthy that prior clinical trials of systemically administered moxonidine as an antihypertensive agent show that moxonidine is well-tolerated. 42–46Furthermore, moxonidine presents an improved side-effect profile over clonidine in terms of reduced sedation and dry mouth, 42,43rebound withdrawal syndrome, 1,45,47and hypotensive effects in normotensive subjects. 48The data presented here would predict that moxonidine may prove effective as a spinal antihyperalgesic agent or coadjuvant to morphine for the treatment of neuropathic pain in humans.

The authors thank Drs. Jin Mo Chung and Seo Lee for instruction in the spinal nerve ligation method, Dr. Laura S. Stone for helpful discussions, and Dr. Dieter Ziegler and Solvay Pharma for the gift of moxonidine.