Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The antinociceptive effects of opioids are enhanced by coadministration of alpha-2 adrenergic agonists such as clonidine in mice and rats (Roerig, 1995; Meert and De Krock, 1994). These findings in experimental animals have relevance for pain management in clinical situations. Clonidine reduces postoperative opioid requirements (Park et al., 1996) and enhances morphine-induced analgesia (De Kock et al., 1992). A spinal site of action for the opioid/alpha-2 interaction is likely (Hylden and Wilcox, 1983), but the mechanism for the spinal morphine/clonidine synergism remains unclear.

Both opioid and alpha-2 receptors are expressed in superficial laminae of the dorsal horn of the spinal cord (Arvidsson et al., 1995; Nicholas et al., 1996) where agonist antinociceptive actions are likely to converge. Stimulation of opioid and alpha-2 receptors produces similar effects on signal transduction pathways (reviewed in Ruffolo et al., 1991; Law and Loh, 1992). Agonists of opioid and alpha-2 receptors have been classically shown to inhibit adenylyl cyclase activity and voltage-dependent calcium channel activity. Recent evidence also suggests that activation of opioid and alpha-2 receptors also stimulates phospholipase C activity (Smart and Lambert, 1996; Cotecchia et al., 1990). These actions alter intracellular concentrations of second messengers, cAMP, calcium, inositol 1,4,5-trisphosphate and diacylglycerol. These intracellular messengers alter the activities of protein kinases such as cAMP-dependent protein kinase (PKA) and protein kinase C (PKC) that, in turn, regulate activities of other cellular proteins through phosphorylation. Activation of these second messenger systems likely contributes to the expression of agonist-induced antinociception. It is possible that cross-talk between pathways may provide a mechanism for opioid/alpha-2 synergism.

The role of second messengers in opioid-induced antinociception is not completely resolved. Agents that increase intracellular cAMP concentrations (Nicholson et al., 1991) or stimulate PKC activity (Zhang et al., 1990; Narita et al., 1996, 1997) attenuate spinal opioid-induced antinociception. In contrast, inhibitors of voltage-dependent calcium channel activity enhance the antinociceptive effects of both opioids and clonidine (Wei et al., 1996; Omote et al., 1993). Calcium channel activity is also likely involved in the spinal opioid/alpha-2 synergism since recent findings show that certain calcium channel antagonists change the synergistic interaction to an additive interaction (Wei et al., 1996; Roerig and Howse, 1996). Thus, altering intracellular second messenger concentrations as well as direct stimulation of PKC or PKA activity alters agonist-induced effects and may also affect the interactions between agonists.

To determine whether PKA or PKC activity contributes to spinal opioid/alpha-2 synergism, our studies were performed using agents that alter PKC or PKA activity. The interaction between spinal morphine and clonidine was determined after pretreatment with agents that selectively inhibit PKC or PKA activity. The effects of PKA stimulation, using agents that increase intracellular cAMP, and PKC stimulation, by phorbol ester, were also examined. A role for PKC or PKA in the morphine/clonidine interaction may be inferred if any of these agents disrupt the synergism that is normally observed. Results of these studies suggest that PKC activity may play a larger role in morphine/clonidine synergism than does the activity of PKA.

	Materials and methods

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and drug administration. Male, ICR mice (25-30 g, Harlan Sprague Dawley, Indianapolis, IN) were used in all studies. Drugs used were morphine sulfate (National Institutes on Drug Abuse), clonidine HCl, and PDBu (Sigma Chemical Co, St. Louis, MO), the selective PKC inhibitors chelerythrine chloride and calphostin C (Research Biochemicals International, Natick, MA), forskolin (water soluble form) and the selective PKA inhibitor H-89 (Calbiochem, San Diego, CA). Rolipram was a generous gift from Schering, Berlin, West Germany. Morphine, clonidine and forskolin were dissolved in sterile water and injected i.t. in a volume of 5 µl (Hylden and Wilcox, 1980). Chelerythrine, calphostin C, H-89, PDBu and rolipram were dissolved initially in DMSO and diluted to the final concentration in water before injection, so that the final concentration of DMSO was 10% or less. All doses are expressed as the salt form.

Measurement of antinociception. Antinociception was measured using the radiant heat tail flick test using a pre-drug latency time of 2 to 3 sec and a cut-off time of 10 sec as previously described (Roerig, 1995). For initial time course studies, pretreatment agent or vehicle was administered i.t. and tail flick responses were measured at 10-min intervals for 60 min. For drug studies, pretreatment agent or vehicle (water or 10% DMSO) was administered i.t., and 15 min (for forskolin and rolipram) or 30 min (for chelerythrine, calphostin C, H-89 and PDBu) later morphine, clonidine or morphine plus clonidine combinations were injected i.t. and tail flick response was measured after 10 min. When morphine and clonidine were coadministered, a constant, equi-effect ratio of drugs was administered for determination of the combined drug ED₅₀ values as previously described (Wei et al., 1996). The morphine/clonidine ratio used for each pretreatment agent was determined from the ED₅₀ value determined for each drug, administered alone, after the pretreatment. Eight to 10 mice were used for each drug dose or combination of drugs and at least four drug doses were used for determination of each ED₅₀ value. Each animal was used only once.

Statistical analysis. The % MPE of each drug treatment in each animal was calculated as % MPE = (post-drug time  pre-drug time)/(10  pre-drug time) × 100. The % MPE values were used to compute ED₅₀ values (and 95% confidence intervals) using the Graded Dose Response program of Tallarida and Murray (1987). Interactions between morphine and clonidine were determined isobolographically from comparison of calculated theoretical additive ED₅₀ values with the experimentally derived ED₅₀ values of the coadministered agonists (Tallarida et al., 1989).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

The time course studies showed that i.t. water pretreatment had no effect on tail flick latency times (data not shown). The 10% DMSO slightly increased tail flick latencies 40 min after administration (2.6 ± 0.1 sec preinjection, 3.0 ± 0.2 postinjection). Chelerythrine and calphostin C (50 pmol doses) did not significantly affect tail flick responses (data not shown). The H-89 at both 50 and 500 pmol doses slightly increased the tail flick latencies 40 min after administration, the time of measurement of morphine and clonidine-induced effects after H-89 pretreatment. For the 50 pmol dose the responses were 2.2 ± 0.1 sec before H-89 injection, and 2.9 ± 0.24 postinjection and for the 500 pmol dose, the values were 2.2 ± 0.1 and 3.4 ± 0.2 sec, respectively. These changes in latency times were significantly different from preinjection times for both doses of H-89. A time course showed that similar increases from preinjection latency times also occurred at 10, 20, 30 and 40 min after the kinase inhibitor administration. By 60 min, response latencies had returned to preinjection times. There was no difference in latency response times among the 10, 20 and 30 min time points. Thus, H-89 slightly increased the tail flick response times by 10 min after administration and this increase remained constant for about 40 min (data not shown).

To allow for any effect on morphine and clonidine antinociception that the 10% DMSO could have, ED₅₀ values for morphine, clonidine and morphine plus clonidine were determined in mice pretreated with 10% DMSO or with water. The dose response curves for morphine and clonidine administered separately after DMSO pretreatment are shown in figure 1 and ED₅₀ values calculated from these curves are shown in tables 1 and 2. Comparable values for water pretreatment are shown in table 2. The ED₅₀ values were not different between the two pretreatments. In the remaining studies, ED₅₀ values obtained after pretreatment with kinase modifiers were compared to the appropriate vehicle controls.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Effect of PKC or PKA inhibitors on i.t. morphine and clonidine-induced tail flick antinociception. Dose response curves for morphine (A) and clonidine (B) after pretreatment with DMSO (10%), inhibitors of PKC (chelerythrine or calphostin C) or an inhibitor of PKA (H-89), all in 50 pmol doses. The inhibitors or DMSO were administered i.t. 30 min before i.t. administration of the morphine or clonidine. Points represent mean ± S.E.M. values for 8 to 15 mice.

Preliminary experiments with various doses of chelerythrine and 100 pmol morphine showed that doses of 1, 10, 50, 100 and 1,000 pmol of the PKC inhibitor had no effect on the antinociceptive action of morphine (data not shown). To assure the selectivity of chelerythrine for inhibition of PKC activity, a 50-pmol dose was used in subsequent experiments. Chelerythrine inhibits PKC activity by interacting with the catalytic domain of PKC (Herbert et al., 1990). To test whether inhibition at the PKC regulatory site would affect the morphine/clonidine synergism, another PKC inhibitor, calphostin C was also tested since this agent inhibits the binding of [³H]-PDBu to PKC (Hidaka and Kobayashi, 1992). Dose response curves for both morphine and clonidine after pretreatment with either chelerythrine or calphostin C are shown in figure 1. The ED₅₀ values derived from these curves, listed in table 1, suggest that chelerythrine increased the potency of clonidine (about 2-fold), but had no effect on morphine-induced antinociception. Calphostin C did not alter the ED₅₀ value of either morphine or clonidine. The clonidine dose response curves in figure 1B after chelerythrine and calphostin C pretreatment are nearly superimposible, shifted to the left of the DMSO pretreatment dose response curve and the ED₅₀ values (table 1) were not significantly different. However, ED₅₀ value after calphostin pretreatment was not significantly different from the ED₅₀ value obtained after DMSO pretreatment.

The interaction between morphine and clonidine was synergistic after DMSO and water pretreatments as has been previously reported in untreated animals. The theoretical additive ED₅₀ values for morphine and clonidine obtained were greater than the experimentally determined ED₅₀ values (table 1). When the ED₅₀ values were plotted in an isobolographic format, the experimental values were below the additive interaction line, indicating that the interaction between morphine and clonidine was synergistic (fig. 2A). In chelerythrine or calphostin C-pretreated mice, the experimentally determined ED₅₀ values were not different from the theoretical additive ED₅₀ values (table 1). When these ED₅₀ values were plotted in isobolograms, the experimental values were within the 95% confidence limits of the additive interaction line (fig. 2, C and D), indicating that the interaction between morphine and clonidine was additive.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Effect of PKC or PKA inhibitors on the interaction between i.t. morphine and clonidine in the tail flick test. Isobolograms for interaction of i.t. coinjected morphine and clonidine after treatment with i.t. protein kinase inhibitors. Values on the axes represent ED50 values obtained from the drugs administered separately. The diagonal line connecting the values represents the additive interaction line. The point on this line represents the theoretical additive ED50 values for the coinjected drugs. The point below the line represents the experimentally derived ED50 values. Diagonal lines through the points represent the 95% confidence intervals for the values. A point below the line indicates a synergistic interaction between morphine and clonidine. A point on the additive interaction line indicates an additive interaction between the agonists. A, Isobologram for i.t. morphine and clonidine obtained when mice were pretreated with DMSO. B, Isobologram for i.t. morphine and clonidine after pretreatment with 50 pmol H-89. C, Isobologram for i.t. morphine and clonidine after pretreatment with 50 pmol calphostin C. D, Isobologram for i.t. morphine and clonidine after pretreatment with 50 pmol chelerythrine.

Experiments using the phorbol ester PDBu to stimulate PKC activity were also performed. Results from these experiments are shown in table 1. The phorbol ester slightly, but significantly, attenuated morphine-induced antinociception (about a 2-fold increase in ED₅₀ value), but did not affect the antinociception produced by clonidine. When the agonists were coadministered, the interaction remained synergistic, as shown in table 1 and figure 3C. The theoretical additive ED₅₀ value for morphine was different from the experimentally determined morphine (in combination) ED₅₀ values, as would be expected from a synergistic morphine/clonidine interaction. However, the theoretical additive ED₅₀ value for clonidine was not different from the experimentally determined clonidine (in combination) ED₅₀ value. Thus, although the statistical analysis showed significant synergism, only the morphine component of the drug combination was significantly changed. Examination of the isobologram (fig. 3C) shows that the point for the experimentally determined ED₅₀ values is below the additive interaction line, indicating a synergistic interaction.

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Effect of PKA and PKC activators on the interaction between i.t. morphine and clonidine in the tail flick test. Isobolograms for interaction of i.t. coinjected morphine and clonidine after treatment with i.t. protein kinase stimulators. Graphs are constructed as described for figure 2. A, Isobologram for i.t. morphine and clonidine after pretreatment with 10 µg forskolin. B, Isobologram for i.t. morphine and clonidine after pretreatment with 1 µg rolipram. C, Isobologram for i.t. morphine and clonidine after pretreatment with 50 pmol PDBu.

To examine the involvement of PKA in the morphine/clonidine interaction, the PKA inhibitor H-89 was tested in a 50 pmol pretreatment dose. This dose did not alter the ED₅₀ value of either morphine or clonidine (fig. 1; table 2). Also, the synergistic interaction between morphine and clonidine was not altered by the PKA inhibitor (fig. 2B; table 2). A higher dose of H-89 (500 pmol) was also tested and results shown in table 2 suggest that increasing the dose 10-fold also did not affect the morphine/clonidine synergism. This higher dose also enhanced the antinociceptive effect of clonidine about 2-fold, but did not affect the morphine ED₅₀ value (table 2).

The effects of stimulation of PKA activity were examined using two agents that increase intracellular cAMP. Forskolin directly stimulates adenylyl cyclase activity and rolipram is a specific inhibitor of cAMP-dependent phosphodiesterase activity. Pretreatment with either of these agents attenuated the antinociception produced by morphine (table 2). Morphine ED₅₀ values increased 10.5- and 7.6-fold after forskolin or rolipram, respectively. Clonidine ED₅₀ values increased 5-fold after forskolin pretreatment and 2-fold after rolipram pretreatment. The synergistic interaction between morphine and clonidine was not changed by either forskolin or rolipram (fig. 3, A and B; table 2).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Our studies suggest that PKC activity plays a role in the antinociceptive synergistic interaction between spinal morphine and clonidine. When PKC activity was inhibited at either the catalytic or the regulatory site (with chelerythrine or calphostin C, respectively) the interaction between spinal morphine and clonidine was reduced from synergistic to additive. The morphine/clonidine synergism does not require full activity of all protein kinases because inhibition of PKA activity did not alter the synergism. Stimulation of PKC or PKA activity did not alter the spinal morphine/clonidine synergism, suggesting that the interaction occurs in the presence of activation of both kinases.

PKC and PKA appear to have different roles in mediating morphine or clonidine antinociception versus mediating morphine/clonidine antinociceptive synergism. A role for both PKC and PKA in spinal opioid-induced antinociception is suggested from present results (tables 1 and 2) showing that i.t. morphine-induced antinociception was attenuated by i.t. pretreatment with PDBu, forskolin or rolipram. Other investigators have reported similar results for opioid agonists (Zhang et al., 1990; Narita et al., 1996, Nicholson et al., 1991). For alpha-2 agonists, results are not as consistent. Present studies show that PDBu did not affect clonidine-induced antinociception, although both forskolin and rolipram attenuated clonidine effects. Others have shown that forskolin has no effect on antinociception produced by spinally administered alpha-2 agonists UK14,304 and guanfacine (Uhlen et al., 1990). These different results could be due to the agonist characteristics of the different alpha-2 agonists. Because the protein kinase stimulaters had greater effects on morphine ED₅₀ values than on clonidine ED₅₀ values, it appeared that opioid-induced antinociception was more sensitive to increases in both PKC and PKA activity than was clonidine-induced antinociception.

The mechanism by which protein kinase stimulation blocks agonist-induced antinociception may be deduced from in vitro studies. Phorbol ester treatment of spinal cord-DRG cocultured neurons abolishes the inhibition of calcium influx produced by opioid and alpha-2 agonists (Attali et al., 1991). In NG108-15 neuroblastoma × glioma hybrid cells, phorbol ester treatment attenuates opioid and alpha-2 agonist-induced inhibition of adenylyl cyclase activity (Louie et al., 1990). In Xenopus oocytes expressing cloned opioid receptors, phorbol ester treatment induces desensitization that is reversed by calphostin C (Ueda et al., 1995). Desensitization of electrophysiologic response to opioid agonist in hippocampal slices is temporally related to changes in phosphorylation state of the receptor (Appleyard et al., 1997). Thus, the observed effects of protein kinase stimulation on agonist-induced antinociception may be related to desensitization induced by phosphorylation.

Our studies suggest that PKC inhibition by chelerythrine, but not PKA inhibition (except at a nonselective dose, see below) enhanced clonidine, but not morphine-induced antinociception. In agreement with the results for morphine, others have shown that neither calphostin C nor the PKA inhibitor KT5720 alters antinociception produced by opioids (Narita et al., 1995, 1996; Bernstein and Welch, 1997), nor do the nonspecific protein kinase inhibitors H7 and H8 (Bilsky et al., 1996). The 2-fold enhancement of clonidine antinociception by chelerythrine, but not calphostin C, suggests that the alpha-2 antinociceptive pathway may be more sensitive to regulation by PKC inhibition at the catalytic site, than at the phorbol ester binding (regulatory) site. But because the clonidine ED₅₀ values after pretreatment with chelerythrine and calphostin C did not differ, these results should be interpreted with caution. It does appear that the alpha-2 pathway is more sensitive than the opioid antinociceptive pathway to inhibition of PKC activity. The slight enhancement of clonidine activity by the 500 pmol dose of H-89 may be due to the slight, but significant, increase in tail flick latency time produced by the 500 pmol dose of H-89 given alone. However, the 50 pmol H-89 dose that also slightly increased latency times did not affect the clonidine ED₅₀ value. If inhibition of adenylyl cyclase activity by opioids or alpha-2 agonists is a factor in the antinociceptive response, it is possible that inhibition of PKA activity by other agents (such as H-89) also produces antinociception. Thus, the H-89 response may simply be adding to the clonidine response. Yet because the 50 pmol dose did not alter the clonidine ED₅₀ value, or the morphine ED₅₀ value, this possibility does not seem adequate to explain the results.

Other possible explanations for differences in effect of H-89 on morphine and clonidine antinociception include differential sensitivities of the opioid and alpha-2 systems and nonselectivity of the H-89 dose. In preliminary studies using 1 nmol H-89, ED₅₀ values for both morphine and clonidine were significantly decreased, although the decrease for clonidine (7-fold) was greater than that for morphine (3.6-fold) (Roerig, unpublished results) suggesting that the alpha-2 antinociceptive pathway was more sensitive to PKA inhibition than the opioid pathway. Also, higher doses of H-89 are likely less selective for PKA. The in vitro K_i value for H-89 is around 50 nM (Chijiwa et al., 1990) and the 0.5-nmol dose used in the present studies represents a concentration of 0.1 mM. At this concentration H-89 likely inhibits PKC (K_i = 32 µM) and activity of other protein kinases as well as PKA. Taking into account that chelerythrine decreased the clonidine, but not the morphine ED₅₀ value (table 1), the observed changes induced by H-89 may be due to loss of selectivity of high doses of H-89 as well as to different sensitivities of clonidine and morphine to the effects of PKC inhibition.

In contrast to the probable roles for both PKC and PKA in agonist-induced antinociception, present studies suggest that PKC, but not PKA, activity is essential for morphine/clonidine antinociceptive synergism. The PKC kinase inhibitors, but not PKC activators or modulators of PKA activity decreased the morphine/clonidine synergism to addition. Interestingly, the PKC inhibitors produced the same effect on the morphine/clonidine synergism as does chronic morphine treatment (Roerig, 1995), a change to an additive interaction. These observations indirectly suggest that chronic treatment with morphine may decrease the activity of PKC. Indeed, chronic i.c.v. infusion of chelerythrine attenuates development of physical dependence to morphine (Fundytus and Coderre, 1996). Chronic opioid treatment also decreases immunoreactive PKC expression in both rat and human cerebral cortex (Busquets et al., 1995). In contrast, immunocyotchemical evidence suggests that chronic morphine increases immunoreactive PKC in rat spinal dorsal horn neurons (Mao et al., 1995) as well as increasing binding of [³H] phorbol ester (Mayer et al., 1995). Additional experiments are required in order to determine whether the changes in PKC expression or activity contribute to development of tolerance to morphine.

A role for PKC, but not PKA, in development of opioid tolerance is also implied from studies in which a single i.t. pretreatment with calphostin C, but not KT5720, blocks development of acute opioid tolerance (Narita et al., 1995, 1996). The lack of effect of KT5720 on acute tolerance may relate to the present observations that modulation of PKA activity did not alter the morphine/clonidine interaction (table 2). This lack of effect of PKA modifiers on morphine/clonidine synergism was somewhat surprising because a great deal of evidence suggests that regulation of the cAMP pathway in the brain is an important component of development of opioid tolerance and physical dependence (Nestler and Aghajanian, 1997). It is possible that the cAMP system is differentially affected by chronic opioid treatment in different regions of the central nervous system, brain vs. spinal cord. Thus, either the change in spinal morphine/clonidine synergism is not a mechanism for development of tolerance, or PKA activation is not a component of the spinal opioid/alpha-2 interaction mechanism.

More than one protein kinase probably participates in agonist-induced desensitization and receptor phosphorylation, phenomena likely related to development of tolerance in animal models. In NG108-15 hybrid cells, opioid agonist-induced desensitization is almost abolished by treatment with staurosporine, a nonspecific protein kinase inhibitor, but only partially attenuated by calphostin C (Cai et al., 1996). Staurosporine also blocks phorbol ester induced opioid receptor phosphorylation, but not opioid-induced receptor phosphorylation (Zhang et al., 1996) or desensitization (Kovoor et al., 1995) in Xenopus oocytes. Activation of both PKC and calcium/calmodulin-dependent protein kinase potententiates opioid receptor desensitization in cells expressing cloned opioid receptors (Mestek et al., 1995). In a similar model, PKC depletion does not alter agonist-induced receptor phosphorylation, but overexpression of beta adrenergic receptor kinase enhances both receptor phosphorylation and desensitization (Pei et al., 1995).

Opioid and alpha-2 agonist-induced antinociception and antinociceptive synergism are likely due to the integration of agonist-induced effects on second messenger systems. The role of protein kinases in opioid-induced antinociception is only beginning to be understood and less is known about these enzymes in the alpha-2 receptor-stimulated analgesia pathway or in the antinociceptive interactions between agonists. Our studies suggest that PKC, but not PKA, activity is likely involved in the spinal morphine/clonidine antinociceptive synergism. Further studies are required to determine whether other kinases may also be involved. The relevant substrates for these kinases must also be identified to fully understand the mechanism of synergism.