Synergistic Antinociceptive Interactions of Morphine and Clonidine in Rats with Nerve-ligation Injury

Male Sprague-Dawley rats, weighing 200–350 g, were used in all experiments. These experiments were approved by the Institutional Animal Care and Use Committee. Rats were anesthetized with 2% halothane in oxygen delivered at 2 l/min. The L5/L6 nerve ligation was performed according to the method described by Kim and Chung. [1]The skin was incised over the caudal lumbar region, and the muscles were retracted. The L5 and L6 branches were identified and carefully isolated from the surrounding fascia. A “finger tight” ligature was made with 4–0 silk suture around each branch proximal to the confluence into the left common sciatic nerve. Hemostasis was ensured, and the incision was sutured. Sham control rats were treated the same way, except the nerves were not ligated.

During anesthesia, catheters were implanted intrathecally for administration of drugs into the region of the lumbar cord, according to the method described by Yaksh and Rudy. [35]An 8-cm length of PE-10 polyethylene tubing was inserted through an incision made in the atlanto-occipital membrane to the level of the lumbar enlargement. The catheter was secured to the musculature at the incision, which was then closed. The rats received 4.4 mg/kg gentamycin intramuscularly and were allowed 5 days recovery before experimentation began. Intrathecally administered drugs were given in a volume of 5 micro liter, followed by a 9-micro liter flush. Rats received increasing doses of either morphine, clonidine, or a 1:3 fixed ratio of morphine to clonidine by intrathecal injection (n = 5–10 rats per group). Separate groups of nerve-ligated and sham-operated rats received intrathecal doses of [D-Ala², NMePhe⁴, Gly-ol]enkephalin (DAMGO) to explore the possible loss of antinociceptive potency of a high-efficacy opioid micro agonist in nerve-injured rats.

Nociceptive testing was performed by the 55 degrees Celsius hot water tail-flick test. The latency to withdrawal of the tail from a water bath maintained at 55 degrees Celsius was determined once before and at 15-min intervals after intrathecal drug administration. A cut-off latency of 10 s was used to prevent tissue injury. Increasing doses of morphine were administered to construct dose-response curves from data gathered at the time of peak effect; each animal received only a single dose of morphine. Naloxone (5 mg/kg, intraperitoneally) was administered 1 h after drug administration to determine whether the effects were mediated via opioid receptors. Data were converted to %MPE (maximal possible effect) by the formula:%MPE = 100 x (test latency-control latency)/(10-control latency). The observer could not be blinded to the condition of the rats (nerve-ligated or sham-operated) because rats with the nerve-ligated revealed their condition by guarding the injured paw. Although this situation may allow some bias on the part of the experimenter, this was guarded against by only measuring one tail-flick response at each time point. The tail-flick response is a clear, nonarbitrary response on the part of the rat to the thermal stimulus.

The interaction between clonidine and morphine was examined by isobolographic analysis, described in detail by Tallarida and colleagues. [36–38]Dose-effect curves were constructed for intrathecal morphine, clonidine, and a 1:3 fixed ratio of morphine to clonidine for L5/L6 nerve-ligated and sham-operated rats. For each drug or drug combination, the A⁵⁰(i.e., dose calculated to produce 50 %MPE) and its associated variances and 95% confidence intervals (95% CI) in terms of total dose were calculated from the log-dose-response curves. The confidence intervals were then arithmetically arranged around the A⁵⁰by the formula In(10) x A⁵⁰x standard error of log A⁵⁰. This transformation is necessary for subsequent statistical calculations of the isobolographic analysis, and the derivation of these calculations are described in detail by Tallarida and colleagues. [36,38]Where only a single drug was given, the “total dose” was the dose of the drug alone; otherwise, it represented the amount of clonidine plus morphine given. A theoretical additive A⁵⁰is calculated for the drug combination. If the relative potency of the additive A⁵⁰relative to the experimentally derived A⁵⁰for the mixture is significantly (P less or equal to 0.05) greater than 1, a synergistic interaction is indicated. An additive interaction is indicated when the theoretical and experimental A⁵⁰values were not significantly different from each other. When clonidine was administered intrathecally to nerve-ligated rats, no A⁵⁰value could be calculated. In that instance, the isobolographic method described by Porreca and colleagues [39]to determine synergistic interaction when one component of the drug mixture is inactive was applied.

Posttreatment means for each time-effect curve were compared with baseline tail-flick latencies by analysis of variance, followed by post hoc analysis (least significance difference test). A probability level of 0.05 indicated significance.

Separate groups of sham-operated and nerve-ligated animals received 4 mg/kg phentolamine intraperitoneally 20 min before the intrathecal administration of 30 micro gram morphine.

The mean baseline tail-flick latency of sham-operated rats was 2.93 +/- 0.10 s, and that of nerve ligated rats was 3.15 +/- 0.09 s. These baseline latencies did not differ between the groups. Under the conditions of the current experiments, it was not possible to detect hyperalgesia of the tail.

Both morphine and clonidine produced dose-dependent antinociception. Morphine produced a maximal effect of 100 +/- 0 %MPE at the largest dose tested, 60 micro gram (Figure 1). The A⁵⁰of morphine was 15 +/- 4.9 micro gram (Table 1). No behavioral deficits were observed at these doses of morphine. Clonidine produced a maximal effect of 88 +/- 9.5 %MPE (Figure 1). Animals that received the highest dose of clonidine, 200 micro gram, appeared mildly sedated, but there was no impairment of the ability of the animal to produce a vigorous tail-flick when the nociceptive stimulus was applied. The calculated A⁵⁰of clonidine was 120 +/- 24 micro gram (Table 1). Administration of a 1:3 fixed ratio of morphine to clonidine produced dose-dependent antinociception (Figure 2), with a maximal effect of 100 +/- 0 %MPE at the largest dose tested (130 micro gram total dose;Figure 1). The A⁵⁰(+/- 95% CI) of the mixture was 15 +/- 9.2 micro gram (Table 1). This value was significantly less than the calculated theoretical additive A sub 50 (+/- 95% CI) of 44 +/- 10 micro gram, which indicated a synergistic antinociceptive interaction between morphine and clonidine.

In L5/L6 nerve-ligated rats, morphine also produced a dose-dependent antinociceptive effect, although to a lesser extent than in the sham-operated animals. The largest dose tested, 100 micro gram, produced a maximal effect of 69 +/- 21.9 %MPE (Figure 3). The A⁵⁰(+/- 95% CI) for morphine was 30 +/- 18 micro gram (Table 1). Clonidine produced a limited dose-dependent effect, with a plateau at the 100- and 200-micro gram doses. The maximal responses observed at the latter doses were 55 +/- 10.9 and 49 +/- 10.2 %MPE (Figure 3). To confirm that a plateau in the dose-effect curve of clonidine had been reached, an additional group of rats received a dose of 400 micro gram intrathecally, which produced a maximal effect of 49 +/- 9.3 %MPE. Although sedation was clearly evident in these animals, the rats were still able to produce a vigorous tail-flick response when the nociceptive stimulus was applied, suggesting that the ability to respond to the thermal stimulus was not impaired. Because of this flattening of the dose-effect curve, it was not possible to calculate an A⁵⁰value for clonidine in the nerve-ligated rats. In contrast to morphine or clonidine alone, the 1:3 mixture of morphine and clonidine produced definitive dose-dependent antinociceptive effects (Figure 4). The dose-effect curve of the morphine-clonidine mixture was shifted to the left of the dose-effect curves for either drug administered alone (Figure 3). The calculated A⁵⁰(+/- 95% CI) for the mixture, given as total dose, was 11 +/- 5.4 micro gram (Table 1). The calculated theoretical additive A⁵⁰(+/- 95% CI) was found to be 118 +/- 73 micro gram (Table 1), which was significantly greater than the experimental A⁵⁰, which indicated a synergistic antinociceptive interaction.

The intraperitoneal administration of 4 mg/kg phentolamine alone did not produce any changes in tail-flick latency in either nerve-ligated or sham-operated animals (data not shown). The maximal effect of 30 micro gram intrathecal morphine observed after phentolamine administration was 34 +/- 2.9 %MPE in sham-operated rats and 18 +/- 4 %MPE in L5/L6 ligated rats (Figure 5). These values were significantly less than the maximal effects observed after morphine alone, 60 +/- 14.4 %MPE and 35 +/- 7.5 %MPE for sham-operated and nerve-ligated rats, respectively (Figure 5). Like morphine, intrathecal DAMGO also produced dose-dependent antinociception in the sham-operated and nerve-injured rats, along with a significant fourfold shift to the right of the dose-response curve (Figure 6). The A⁵⁰values (95% CI) for DAMGO in the sham-operated and ligated groups were 0.23 micro gram (range, 0.088–0.60) and 0.97 micro gram (range, 0.34–2.7), respectively. Unlike morphine, however, there was no loss of efficacy for intrathecal DAMGO in the nerve-injured rats; a full effect of 100 +/- 0 %MPE was observed with 10 micro gram DAMGO (Figure 6).

The observation that the intrathecal administration of morphine and clonidine produced a synergistic effect with regard to a spinally mediated nociceptive response, the tail-flick reflex, in sham-operated rats is in accordance with the established literature. A supra-additivity was described between morphine and clonidine in several species and after both systemic and intrathecal routes of administration. [21,31–34]Therefore, this observation establishes the validity of the methods used in the current study.

The acute antinociceptive action of clonidine was clearly diminished in potency and efficacy after nerve ligation injury. Noteworthy, however, is the unexpected loss of antinociceptive activity of clonidine, despite the observed and established antiallodynic action of lower (i.e., 20–60 micro gram) doses of this compound. [6,11]The test dose of 400 micro gram intrathecally confirmed that a plateau in the dose-effect curve of clonidine occurred. Although sedation was evident in these animals, the rats still produced a vigorous tail-flick response when the nociceptive stimulus was applied, indicating that the sedation did not interfere with the animals' ability to respond to the thermal stimulus. In other studies, 100 micro gram intrathecal clonidine produced a maximal effect in the hot plate test at 52 degrees Celsius, [40]and 750 micro gram/kg clonidine subcutaneously produced a near maximal effect in a radiant heat tail-flick test. [41]The doses used and responses obtained for sham-operated animals in this study were well within the equivalent range indicated by these studies.

Suggestions of considerable overlap between mechanisms involved in the development of hyperalgesia subsequent to nerve injury and mechanisms observed after development of morphine tolerance have been made. [42,43]Both hyperalgesia and morphine tolerance, which may be associated with activation of N-methyl-D-aspartate receptors and increased activity of intracellular protein kinase C, were shown to be prevented by pretreatment with the N-methyl-D-aspartate antagonist MK-801. [42–44]These observations were taken to indicate that common neural substrates exist between morphine tolerance and loss of morphine activity in the presence of nerve injury, [42,43]and may explain the loss of antinociceptive efficacy of morphine after nerve injury seen in the current study and reported previously. [13] 

Although similar mechanisms may be invoked, we must consider the possibility that morphine tolerance and loss of morphine activity after nerve injury are not identical phenomena. In the current study, both morphine, an opioid micro receptor agonist, and clonidine, an alpha²adrenergic receptor agonist, demonstrated significant loss of activity after nerve ligation injury. However, there is little evidence of antinociceptive cross-tolerance between morphine and clonidine. The antinociceptive effect of systemically or intrathecally administered morphine was virtually abolished in mice made tolerant to morphine, whereas that of clonidine administered systematically or intrathecally was not altered, [20,30]a finding consistent with actions of these compounds at opioid micro and adrenergic alpha²receptors, respectively. In addition, it was reported that a single administration of MK-801 does not reverse morphine tolerance, [45]whereas it restores the antinociceptive efficacy of intrathecal morphine in the nerve-ligated rat. [12]Likewise, a single pretreatment with MK-801 also restores the antiallodynic activity of morphine, even though neither compound demonstrates antiallodynic activity alone in this ligation model.*,**

The reasons for the loss of antinociceptive activity of morphine and clonidine in the nerve-ligated rat are unclear and will require further investigation, but speculation about the mechanism is possible. Nerve ligation injury may hypersensitize the spinal cord to nociceptive stimuli, such that nociceptive processing is augmented as if the intensity of the stimuli were greater than in the nonligated rat, thereby increasing the opioid receptor occupancy requirement. However, it was shown recently that even when low levels of stimulus intensities were applied (i.e., 48 degrees Celsius warm water or low-intensity radiant heat tail-flick tests), there was no evidence of hyperalgesia in the nerve-ligated rat.** In addition, the observation that DAMGO, a high-efficacy opioid micro agonist that consequently has a lower receptor occupancy requirement than morphine, [46,47]also lost potency after nerve ligation injury also argues against a greater receptor occupancy requirement as a mechanism by which intrathecal opioids lose antinociceptive activity after peripheral nerve injury. An important consideration in terms of an antinociceptive action, however, is sustained afferent drive resulting from ectopic foci associated with the nerve injury site, which may affect the ability of morphine to produce a significant antinociceptive [12]or antiallodynic response.** Therefore, blockade of N-methyl-D-aspartate receptors with MK-801 or local application of bupivacaine at the site of the injury restores the efficacy of morphine, [12]and such a situation may equally apply to the antinociceptive actions of clonidine. In addition, the possibility also exists that changes in neurotransmitter activity that act as “anti-analgesic”(e.g., cholecystokinin) agents may contribute to the loss of clonidine antinociceptive efficacy, as occurs with morphine, though no evidence currently supports this possibility.

These findings may also offer important mechanistic insights into the underlying cause of the allodynia resulting from nerve injury. After peripheral nerve injury, there is considerable sympathetic innervation of the dorsal root ganglia, which may maintain the allodynic state in this model. [14,17]It was clearly demonstrated that activation of alpha²-adrenergic receptors of preganglionic sympathetic neurons results in a hyperpolarization of these cells in culture, [48]and that clonidine inhibits the activity of the preganglionic sympathetic neurons. [49]The intrathecal administration of low doses of clonidine to rats produced significant reductions in firing rates of neuronal activity of the lumbar sympathetic chain and also altered thermoregulation [50]and reduced resting blood pressure and heart rate, [51]which indicates a reduction in efferent sympathetic activity. Clonidine, administered intrathecally, at lower doses than used in these studies, attenuated tactile allodynia in the L5/L6 nerve ligation model, and this effect was attributed to reduced spinal sympathetic outflow from preganglionic sympathetic neurons. [6]In addition, it also was shown that bilateral lumbar sympathectomy attenuated allodynia in rats with L5/L6 nerve ligation injury, and that the allodynia was rekindled by the administration of norepinephrine to the previously allodynic site, strongly suggesting that a sympathetic component exists in this model of neuropathic pain. [19]Finally, clinically, a sympathetic component of neuropathic pain is also suggested by the observations that chemical sympathectomy, by administration of clonidine [52]or phentolamine, [53]effectively diminishes painful neuropathy in humans. For this reason, the antiallodynic actions of clonidine have been thought to be possibly related to an effect on afferent input or, alternatively, to be the result of a sympatholytic action. The data presented here that show that clonidine is only partially active in inhibiting the tail-flick response at doses that are fully active against allodynia strongly suggest that the observed antiallodynic action [6]occurs by blocking spinal sympathetic outflow and not by acting to limit afferent nociceptive input. For these reasons, clonidine can be demonstrated to be fully antiallodynic, but only partially effective in suppressing acute nociceptive stimuli, as shown in the tail-flick test in nerve-injured animals. Again, the partial effectiveness of clonidine as an antinociceptive agent in nerve-injured animals may reflect changes in responsiveness in the spinal cord due to sustained afferent drive or alterations in the levels of neurotransmitters with opposing actions.

Finally, the results reported here are particularly noteworthy because clonidine was virtually inactive as an antinociceptive agent in the animals with nerve injury, yet produced a robust antinociceptive synergistic interaction with morphine in this model of neuropathic pain. In fact, the potencies of the morphine-clonidine mixtures were not different from each other in the L5/L6 ligated or sham-operated animals. Therefore, we may speculate that the antinociceptive effect observed in these nerve-injured animals between morphine and clonidine is not the result of the sympatholytic effect of clonidine, but of other mechanisms that are likely to be occurring in the spinal cord. It might be suggested that the potentiation of morphine antinociception was, in fact, dependent on a blockade sympathetic outflow elicited by intrathecal clonidine, possibly resulting in a reduction of spontaneous pain secondary to the nerve ligation injury. In that case, however, phentolamine, which produces a sympatholytic effect equivalent to sympathetic ganglion block by local anesthesia, [54]would be expected to augment the antinociceptive effect of morphine in the current study. However, the observation that phentolamine appeared to diminish the antinociceptive effect of morphine in both the nerve-ligated and sham-operated rats suggests that thermal hyperalgesia, unlike tactile allodynia, may not be sympathetically dependent. Also note that phentolamine was demonstrated to attenuate the antinociceptive effect of morphine administered either systemically or centrally in healthy animals. [54,55] 

In summary, the data presented in this study strongly indicate that the antinociceptive effect of both morphine and clonidine are significantly diminished after peripheral nerve injury. This loss of efficacy may be attributed to neuroplastic changes, including changes in levels of endogenous antianalgesic neurotransmitters as well as sustained afferent drive. Noteworhty, however, is the observation that the effective doses of the combination of morphine and clonidine were significantly reduced, yet were fully efficacious as antinociceptive agents in this model of neuropathic pain. Combination therapy involving spinal administration of mixtures of clonidine and morphine, known to produce relief against acute pain clinically, may be of significant clinical importance in the treatment of patients suffering from neuropathic states.

*Nichols ML, Bian D, Lopez Y, Ossipov MH, Porreca F: Restoration of the antiallodynic and antinociceptive efficacy of morphine by antisera to dynorphin A(1–13) or MK801 in a nerve-ligation model of peripheral neuropathy (abstract). VIIIth World Congress on Pain 1996; 351.

**Bian D, Wgert S, Ossipov MH, Vanderah T, Nichols ML, Porreca F: Differential sensitivities of hyperesthesias after L5/L6 nerve ligation injury to spinal pharmacologic manipulations (abstract). VIIIth World Congress on Pain 1996; 350.