Multiple Receptors Involved in Peripheral 2, , and a Antinociception, Tolerance, and Withdrawal

The Journal of Neuroscience, January 15, 1997, 17(2):735–744

Multiple Receptors Involved in Peripheral ␣2, ␮, and A1 Antinociception, Tolerance, and Withdrawal

K. O. Aley and Jon D. Levine Departments of Anatomy, Medicine, and Oral and Maxillofacial Surgery, Division of Neuroscience and Biomedical Sciences Program, University of California at San Francisco, San Francisco, California 94143-0452

We examined the interactions among three classes of not only by yohimbine (␣2-antagonist) but also by PACPX (A1- peripherally-acting antinociceptive agents (␮-opioid, ␣2- antagonist) and by naloxone (␮-antagonist), and that DAMGO adrenergic, and A1-adenosine) in the development of tolerance (␮-agonist) antinociception and CPA (A1-agonist) antinocicep- and dependence to their antinociceptive effects. Antinocicep- tion were blocked not only by naloxone (␮-antagonist) and tion was determined by assessing the degree of inhibition of PACPX (A1-antagonist), respectively, but also by yohimbine prostaglandin E2 (PGE2)-induced mechanical hyperalgesia, us- (␣2-antagonist). This cross-antagonism of antinociception oc- ing the Randall-Selitto paw-withdrawal test. curred at the ID80 dose for each antagonist at its homologous Tolerance developed within 4 hr to the antinociceptive effect receptor. To test the hypothesis that the physical presence of of the ␣2-adrenergic agonist clonidine; dependence also oc- ␮-opioid receptor is required not only for ␮ antinociception but curred at that time, demonstrated as a withdrawal hyperalgesia also for ␣2 antinociception, antisense oligodeoxynucleotides that was precipitated by the ␣2-receptor antagonist yohimbine. (ODNs) for the ␮-opioid and ␣2C-adrenergic receptors were These ﬁndings are similar to those reported previously for administered intrathecally to reduce the expression of these tolerance and dependence to ␮ and A1 peripheral antinocicep- receptors on primary afferent neurons. These studies demon- tion (Aley et al., 1995). strated that ␮-opioid ODN administration decreased not only

Furthermore, cross-tolerance and cross-withdrawal between ␮-opioid but also ␣2-adrenergic antinociception; A1 antinoci- ␮,A1, and ␣2 agonists occurred. The observations of cross- ception was unaffected. In contrast, ␣2C-adrenergic ODN de- tolerance and cross-withdrawal suggest that all three receptors creased antinociception induced by all three classes of antino- are located on the same primary afferent nociceptors. In addi- ciceptive agents. tion, the observations suggest that the mechanisms of toler- In conclusion, these data suggest that peripheral antinoci- ance and dependence to the antinociceptive effects of ␮,A1, ception induced by ␮, ␣2, and A1 agonists requires the physical and ␣2 are mediated by a common mechanism. presence of multiple receptors. We propose that there is a ␮, Although any of the agonists administered alone produce A1, ␣2 receptor complex mediating antinociception in the pe- antinociception, we found that ␮,A1, and ␣2 receptors may not riphery. In addition, there is cross-tolerance and cross- act independently to produce antinociception, but rather may dependence between ␮,A1, and ␣2 antinociception, suggest- require the physical presence of the other receptors to produce ing that their underlying mechanisms are related. antinociception by any one agonist. This was suggested by the Key words: pain; analgesia; dorsal root ganglion; opioid;

ﬁnding that clonidine (␣2-agonist) antinociception was blocked antisense oligodeoxynucleotide; receptor cross-talk

Both ␮-opioid and A1-adenosine agonists have been shown to ␣2-adrenergic agonists have also been shown to produce an- produce a potent antinociception when administered in the pe- tinociception when administered peripherally (Khasar et al., riphery (Taiwo and Levine, 1990; Aley et al., 1995). This antino- 1995). Because many of the effects of ␣2 agonists, like those ciception is detected in the presence of hyperalgesia produced by produced by ␮ and A1 agonists, involve Gi protein signaling numerous inﬂammatory mediators including PGE2 (Taiwo and (Sharma et al., 1975; Law et al., 1981; Childers and LaRiviere, Levine, 1990). Many of the effects produced by ␮ and A1 agonists 1984; Mankman et al., 1988), we hypothesized that they all pro- are mediated through a common second messenger, speciﬁcally, duce antinociception in the periphery through common cellular activation of an inhibitory guanine nucleotide binding (Gi) protein mechanisms. In addition, we hypothesized that there would be (Sharma et al., 1975; Law et al., 1981; Childers and LaRiviere, symmetric cross-tolerance and cross-dependence between the pe- 1984; Mankman et al., 1988). ripheral antinociceptive actions of these three agonists. Because

We have found that tolerance and dependence develop to both ␣2-adrenergic, ␮-opioid, and A1-adenosine agonists are widely ␮ and A1 peripheral antinociception. In addition, a symmetric used clinically, interactions between their peripheral antinocicep- cross-tolerance and cross-dependence exists between the ␮ and tive effects is of signiﬁcant interest.

A1 antinociceptive mechanisms (Aley et al., 1995). MATERIALS AND METHODS

Received Sept. 3, 1996; revised Oct. 22, 1996; accepted Oct. 25, 1996. Animals Correspondence should be addressed to Dr. Jon D. Levine, Department of Experiments were performed on male Sprague Dawley rats (250–300 gm, Anatomy, University of California-San Francisco, 521 Parnassus Avenue, San Fran- Bantin and Kingman, Fremont, CA). Animals were housed in groups of cisco, CA 94143-0452. two under a 12 hr light/dark cycle (light on 6.0 hr). Food and water were Copyright ᭧ 1997 Society for Neuroscience 0270-6474/97/170735-10$05.00/0 available ad libitum. All testing was done between 10.0 and 16.0 hr. 736 J. Neurosci., January 15, 1997, 17(2):735–744 Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia

Table 1. Abbreviations of agents used the drugs except the oligodeoxynucleotides were administered intrader- mally in a volume of 2.5 ␮l/paw. Abbreviation(s) Agent Action Intrathecal cannulation

PGE2 or E2 Prostaglandin E2 Hyperalgesic inflamma- To administer drugs intrathecally, a catheter (PE-10 polyethylene tubing) tory mediator was inserted caudally 8.5 cm into the subdural space through a midline DAMGO or D [D-Ala2,N-Me-Phe4,gly5-ol] ␮-opioid receptor agonist incision made in the atlanto-occipital membrane of rats anesthetized with pentobarbital (50 mg/kg, i.p.); the external end of the catheter was enkephalin secured to the skull with screws and dental acrylic (Yaksh and Rudy, N Naloxone Opioid receptor 1976). The skin incision was sutured closed, and the animals were allowed antagonist to recover. Two days after surgery, rats that showed no motor deficits were used for experimental studies involving intrathecal administration of Cl Clonidine ␣2-adrenergic agonist ODNs. Yo Yohimbine ␣2-adrenergic antagonist 6 CPA N -cyclopentyl-adenosine A1-adenosine receptor Antisense oligodeoxynucleotides agonist ␮-Opioid receptor oligodeoxynucleotides. The ␮-opioid receptor antisense PACPX 1,3-dipropyl-8-(2-amino-4- A1-adenosine receptor and sense ODNs used in this study were synthesized using a Nucleic Acid chlorophenyl)-xanthine antagonist Synthesizer model 391 (PCR Mate; Applied Biosystems, Foster City, ␮-antisense ␮-opioid receptor Downregulation of CA). The ␮-opioid receptor antisense ODN, 5Ј-CGCCCCAGCC- TCTTCCTCT-3Ј, is directed against the 5Ј-untranslated region of ␮ oligodeoxynucleotide -opioid receptors ␮-opioid receptor-1 (MOR-1) clone, located between bases 87 and 69 (ODN) upstream from the initiating ATG. The sense ODN, 5Ј-AGAGG- ␮-sense No effect AAGAGGCTGGGGCG-3Ј (Rossi et al., 1994), is complementary to the antisense sequence. Concentrations of ODN stocks were determined by ␣2-antisense ␣2C-adrenergic receptor Downregulation of oligodeoxynucleotide ␣ -adrenergic spectrophotometry. Before their use, ODNs were lyophilized and resus- 2C pended in 0.9% NaCl to a concentration of 1 ␮g/10 ␮l. (ODN) receptors Rats were divided into three groups: one group was untreated (without ␣2-sense No effect cannulae), a second group was treated with sense ODN (1 ␮g), and the third group was treated with antisense ODN (1 ␮g/rat). Using a microsy- ringe (Hamilton, Nevada City, UT), a dose of 1 ␮g ODN was administered to each rat intrathecally, in a volume of 10 ␮l, followed by 10 ␮lof Experiments were carried out under approval of the Institutional Animal saline (the dead space of the intrathecal catheter), on alternate days (days Care Committee of the University of California, San Francisco. 1, 3, and 5). Behavioral tests were done 24 hr after the last dose of ODN. We have found that antisense ODN against MOR-1 attenuates ␮-opioid Behavioral testing receptor-like immunoreactivity in the dorsal horn of the spinal cord and The nociceptive flexion reflex was quantified with a Basile Analgesymeter peripheral nerve, DAMGO-induced inhibition of calcium current in (Stoelting, Chicago, IL), which applies a linearly increasing mechanical cultured dorsal root ganglion neurons and DAMGO-induced inhibition

force to the dorsum of the hindpaw of the rat. Before the experiments, of PGE2-induced hyperalgesia (Khasar et al., 1996). rats were exposed to the procedure for3d(1hrdaily at 5 min intervals, ␣2C-opioid receptor oligodeoxynucleotides. We have demonstrated pre- i.e., 12 exposures), a procedure that produces a stable baseline threshold viously that the ␣2-adrenergic receptor mediating peripheral antinocicep- measurement and enhances the ability to detect the action of hyperalgesic tion has the pharmacological characteristics of the ␣2C subtype (Khasar et agents (Taiwo et al., 1989; Aley et al., 1995). On the day of the experi- al., 1995). Therefore, we also synthesized the antisense and sense ODNs

ment, rats were exposed to the same procedure, and the mean of the last for the ␣2C receptor subtype, using Nucleic Acid Synthesizer model 6 of the 12 readings was considered to be the baseline mechanical 391 PCR Mate. The ␣2C receptor antisense ODN, 5Ј-ACCTGC- nociceptive threshold. The mean baseline threshold in these experiments GGAGTACTG-3Ј, was developed by Lingen and Ordway (1995). The was 110.9 Ϯ 0.4 gm (n ϭ 466; mean Ϯ SEM). Mechanical threshold was sense ODN sequence 5Ј-CAGTACTCCGCAGGT-3Ј is complementary again determined at different time points (15, 20, and 25 min) after to the antisense sequence. various treatments. The mean of these three readings was deﬁned as the Rats were divided into three groups: one group was untreated (without paw-withdrawal threshold post-treatment for that paw, and this value was cannulae), a second group was treated with sense ODN (1 ␮g/rat), and used to calculate the percentage change from the baseline threshold [% the third group was treated with antisense ODN (1 ␮g). Treatment and change in threshold ϭ (pretreatment threshold - post-treatment thresh- behavioral testing were as described for ␮-opioid receptor ODN old)/(pretreatment threshold)] ϫ 100. experiments. Abbreviations for the drugs used in this study and their actions are Drug administration shown in Table 1; experimental protocols are shown in Table 2.

The drugs used in this study were PGE2 [100 ng; direct-acting hyperalgesic inﬂammatory mediator (Pitchford and Levine, 1991, Gold et al., Statistical analysis

1994)], DAMGO (␮-opioid receptor agonist), clonidine (␣2-adrenergic Data are presented as mean Ϯ SEM of six or more observations in each receptor agonist), CPA (A1-adenosine receptor agonist), naloxone of the experimental groups. Statistical significance was determined by methyl iodide (opioid receptor antagonist), yohimbine HCl (␣2- ANOVA, followed by Scheffe’s post hoc test; p Ͻ 0.05 was considered adrenergic receptor antagonist), and PACPX (A1-adenosine receptor statistically significant. Some data are repeated for comparison (see figure antagonist), all from Research Biochemicals International (Natick, MA). legends). The selection of the drug doses used in this study was based on the dose–response curves determined during this study or from previous work RESULTS done in this laboratory (Aley et al., 1995). The stock solution of PGE2 (1 Tolerance to clonidine antinociception ␮g/2.5 ␮l) was prepared in 10% ethanol and further dilutions were made in saline; the final concentration of ethanol was Յ1%. DAMGO, In the present study, we found that after repeated administration, clonidine, CPA, naloxone, yohimbine, and PACPX were dissolved in clonidine (100 ng) produces tolerance for its inhibition of PGE2 saline. When drug combinations were used, they were administered from (100 ng)-induced hyperalgesia (Fig. 1A), similar to that seen for ␮ the same syringe so that the drug mentioned first reached the intradermal and A agonists (Aley et al., 1995). site first; the two drugs were separated in the syringe by a small air bubble 1 to avoid the problem of diffusion. When an antagonist was included to Bidirectional cross-tolerance for ␮,A1, and ␣2 antagonize the effect of an agonist, it was always injected first. The ID80 peripheral antinociception dose of each antagonist (i.e., naloxone 200 ng, yohimbine 100 ng, and PACPX 100 ng), calculated from its dose–response curve for reversal of In paws made tolerant to DAMGO or CPA, clonidine failed to the effect of its homologous agonist, was used throughout the study. All produce a significant antinociceptive effect when injected at the Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia J. Neurosci., January 15, 1997, 17(2):735–744 737

Table 2. Experimental protocols

Group N Treatment Dose(s) I–A

1 24 PGE2 100 ng 2 16 Clonidine ϩ PGE2 100 ng ϩ 100 ng 3 6 Clonidine hourly ϫ 3 100 ng ϫ 3

4 12 Clonidine hourly ϫ 3, fourth hour clonidine ϩ PGE2 100 ng ϫ 3, 100 ng ϩ 100 ng I–B

1 12 DAMGO hourly ϫ 3, fourth hour clonidine ϩ PGE2 1 ␮g ϫ 3, 100 ng ϩ 100 ng 2 12 CPA hourly ϫ 3, fourth hour clonidine ϩ PGE2 1 ␮g ϫ 3, 100 ng ϩ 100 ng 3 16 Clonidine ϩ PGE2 100 ng ϩ 100 ng 4 12 Clonidine hourly ϫ 3, fourth hour DAMGO ϩ PGE2 100 ng ϫ 3, 1 ␮g ϩ 100 ng 5 6 DAMGO ϩ PGE2 1 ␮g ϩ 100 ng 6 6 DAMGO hourly ϫ 31␮gϫ3 7 6 CPA hourly ϫ 31␮gϫ3

8 12 Clonidine hourly ϫ 3, fourth hour CPA ϩ PGE2 100 ng ϫ 3, 1 ␮g ϩ 100 ng 9 6 CPA ϩ PGE2 1 ␮g ϩ 100 ng II-A

1 24 PGE2 100 ng

2 16 Clonidine ϩ PGE2 100 ng ϩ 100 ng 3 6 Clonidine ϫ 3 100 ng ϫ 3 4 8 Clonidine hourly ϫ 3, fourth hour yohimbine 100 ng ϫ 3, 100 ng II-B 1 10 Clonidine hourly ϫ 3, fourth hour naloxone 100 ng ϫ 3, 200 ng 2 10 Clonidine hourly ϫ 3, fourth hour PACPX 100 ng ϫ 3, 100 ng 3 6 Clonidine ϫ 3 100 ng ϫ 3 4 10 DAMGO hourly ϫ 3, fourth hour yohimbine 1 ␮g ϫ 3, 100 ng 5 6 DAMGO ϫ 31␮gϫ3 6 10 CPA hourly ϫ 3, fourth hour yohimbine 1 ␮g ϫ 3, 100 ng 7 6 CPA hourly ϫ 31␮gϫ3 III-A

1 24 PGE2 100 ng 2 6 DAMGO ϩ PGE2 1 ␮g ϩ 100 ng 3 6 Naloxone ϩ DAMGO ϩ PGE2 1ngϩ1␮gϩ100 ng 4 6 Naloxone ϩ DAMGO ϩ PGE2 10 ng ϩ 1 ␮g ϩ 100 ng 5 6 Naloxone ϩ DAMGO ϩ PGE2 100 ng ϩ 1 ␮g ϩ 100 ng 6 6 Naloxone ϩ DAMGO ϩ PGE2 1 ␮g ϩ 1 ␮g ϩ 100 ng III-B

1 24 PGE2 100 ng 2 16 Clonidine ϩ PGE2 100 ng ϩ 100 ng 3 6 Yohimbine ϩ clonidine ϩ PGE2 1ngϩ100 ng ϩ 100 ng 4 6 Yohimbine ϩ clonidine ϩ PGE2 10 ng ϩ 100 ng ϩ 100 ng 5 6 Yohimbine ϩ clonidine ϩ PGE2 100 ng ϩ 100 ng ϩ 100 ng 6 6 Yohimbine ϩ clonidine ϩ PGE2 1 ␮g ϩ 100 ng ϩ 100 ng III-C

1 24 PGE2 100 ng 2 6 CPA ϩ PGE2 1 ␮g ϩ 100 ng 3 6 PACPX ϩ CPA ϩ PGE2 1ngϩ1␮gϩ100 ng 4 6 PACPX ϩ CPA ϩ PGE2 10 ng ϩ 1 ␮g ϩ 100 ng 5 6 PACPX ϩ CPA ϩ PGE2 100 ng ϩ 1 ␮g ϩ 100 ng 6 6 PACPX ϩ CPA ϩ PGE2 1 ␮g ϩ 1 ␮g ϩ 100 ng IV-A

1 24 PGE2 100 ng 2 16 Clonidine ϩ PGE2 100 ng ϩ 100 ng 3 6 Yohimbine ϩ clonidine ϩ PGE2 100 ng ϩ 100 ng ϩ 100 ng 4 10 Naloxone ϩ clonidine ϩ PGE2 200 ng ϩ 100 ng ϩ 100 ng 5 10 PACPX ϩ clonidine ϩ PGE2 100 ng ϩ 100 ng ϩ 100 ng IV-B

1 24 PGE2 100 ng 2 6 DAMGO ϩ PGE2 1 ␮g ϩ 100 ng 3 6 Naloxone ϩ DAMGO ϩ PGE2 200 ng ϩ 1 ␮g ϩ 100 ng 4 8 Yohimbine ϩ DAMGO ϩ PGE2 100 ng ϩ 1 ␮g ϩ 100 ng 5 6 PACPX ϩ DAMGO ϩ PGE2 100 ng ϩ 1 ␮g ϩ 100 ng IV-C

1 24 PGE2 100 ng 2 6 CPA ϩ PGE2 1 ␮g ϩ 100 ng

Table 2. continues. 738 J. Neurosci., January 15, 1997, 17(2):735–744 Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia

Table 2. continued

Group N Treatment Dose(s) IV-C

3 6 PACPX ϩ CPA ϩ PGE2 100 ng ϩ 1 ␮g ϩ 100 ng 4 8 Naloxone ϩ CPA ϩ PGE2 200 ng ϩ 1 ␮g ϩ 100 ng 5 6 Yohimbine ϩ CPA ϩ PGE2 100 ng ϩ 1 ␮g ϩ 100 ng V-A 1 8 Clonidine hourly ϫ 3, fourth hour yohimbine 100 ng ϫ 3, 100 ng 2 6 Clonidine hourly ϫ 3, fourth hour clonidine ϩ yohimbine 100 ng ϫ 3, 100 ng ϩ 100 ng 3 8 Clonidine hourly ϫ 3, fourth hour DAMGO ϩ yohimbine 100 ng ϫ 3, 1 ␮g ϩ 100 ng 4 8 Clonidine hourly ϫ 3, fourth hour CPA ϩ yohimbine 100 ng ϫ 3, 1 ␮g ϩ 100 ng V-B 1 6 DAMGO hourly ϫ 3, fourth hour naloxone 1 ␮g ϫ 3, 200 ng 2 6 DAMGO hourly ϫ 3, fourth hour DAMGO ϩ naloxone 1 ␮g ϩ 1 ␮g ϩ 200 ng 3 8 DAMGO hourly ϫ 3, fourth hour clonidine ϩ naloxone 1 ␮g ϩ 100 ng ϩ 200 ng 4 8 DAMGO hourly ϫ 3, fourth hour CPA ϩ naloxone 1 ␮g ϩ 1 ␮g ϩ 200 ng V-C 1 8 CPA hourly ϫ 3, fourth hour PACPX 1 ␮g ϫ 3, 100 ng 2 6 CPA hourly ϫ 3, fourth hour CPA ϩ PACPX 1 ␮g ϩ 1 ␮g ϩ 100 ng 3 8 CPA hourly ϫ 3, fourth hour clonidine ϩ PACPX 1 ␮g ϩ 100 ng ϩ 100 ng 4 8 CPA hourly ϫ 3, fourth hour DAMGO ϩ PACPX 1 ␮g ϩ 1 ␮g ϩ 100 ng VI-A

1 24 PGE2 100 ng 2 6 DAMGO ϩ PGE2 1 ␮g ϩ 100 ng 36␮-antisense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng

DAMGO ϩ PGE2 46␮-sense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng

DAMGO ϩ PGE2 VI-B

1 24 PGE2 100 ng 2 16 Clonidine ϩ PGE2 100 ng ϩ 100 ng 36␮-antisense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 100 ng ϩ 100 ng

clonidine ϩ PGE2 46␮-sense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 100 ng ϩ 100 ng

clonidine ϩ PGE2 VI-C

1 24 PGE2 100 ng 2 6 CPA ϩ PGE2 1 ␮g ϩ 100 ng 36␮-antisense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng

CPA ϩ PGE2 46␮-sense intrathecally alternate days ϫ 3, 24 hr after CPA 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng

ϩ PGE2 VII-A

1 24 PGE2 100 ng 2 6 DAMGO ϩ PGE2 1 ␮g ϩ 100 ng 36␣2-antisense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng

DAMGO ϩ PGE2 46␣2-sense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng

DAMGO ϩ PGE2 VII-B

1 24 PGE2 100 ng 2 16 Clonidine ϩ PGE2 100 ng ϩ 100 ng 36␣2-antisense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 100 ng ϩ 100 ng

clonidine ϩ PGE2 46␣2-sense intrathecally alternate days ϫ 3, 24 hr after 1 ␮g ϫ 3, 100 ng ϩ 100 ng

clonidine ϩ PGE2 VII-C

1 24 PGE2 100 ng 2 6 CPA ϩ PGE2 1 ␮g ϩ 100 ng 36␣2-antisense intrathecally days ϫ 3, 24 hr after CPA ϩ PGE2 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng 46␣2-sense intrathecally alternate days ϫ 3, 24 hr after CPA 1 ␮g ϫ 3, 1 ␮g ϩ 100 ng

ϩ PGE2

2 4 5 Abbreviations: PGE2, Prostaglandin E2 (EP receptor agonist); DAMGO, [D-Ala , N-Me-Phe , gly -ol] (␮-opioid receptor agonist); Cl, clonidine (␣2 agonist); Yo, yohimbine 6 (␣2 antagonist); CPA, N -cyclopentyl adenosine (A1-adenosine agonist); PACPX, 1,3-dipropyl-8-(2-amino-4-chlorophenyl)-xanthine (A1-adenosine antagonist). There are repetitions for the sake of comparison. Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia J. Neurosci., January 15, 1997, 17(2):735–744 739

Figure 1. A, Repeated administration of clonidine produces tolerance to antinociception. Effect of PGE2 (E2), clonidine plus PGE2 (ClϩE2), clonidine once hourly for 3 hr (Clx3), clonidine once hourly for 3 hr, and at the fourth hour clonidine plus PGE2 (Clx3, ClϩE2) on mechanical paw withdrawal threshold in the rat. B, Bidirectional cross-tolerance develops among A1, ␣2, and ␮ antinociception. Effect of clonidine plus PGE2 (ClϩE2), DAMGO once hourly for 3 hr (Dx3), DAMGO once hourly for 3 hr and at the fourth hour clonidine plus PGE2 (Dx3,ClϩE2), CPA once hourly for 3 hr and at the fourth hour clonidine plus PGE2 (CPAx3, ClϩE2), CPA once hourly for 3 hr (CPAx3), clonidine once hourly for 3 hr and at the fourth hour DAMGO plus PGE2 (Clx3, DϩE2), and clonidine once hourly for 3 hr and at the fourth hour CPA plus PGE2 (Clx3, CPAϩE2) on mechanical paw withdrawal threshold in the rat.

fourth hour (Fig. 1B). Also, in paws made tolerant to clonidine, suggest that the ␣2 receptor is involved not only in ␣2 antinoci-

DAMGO and CPA were no longer antinociceptive (Fig. 1B). ception but also in ␮ and A1 antinociception. In addition, the data

These observations suggest that there is bidirectional cross- suggest that the ␮ and A1 receptors are involved in ␣2 tolerance between ␣2, ␮, and A1 to their peripheral antinocicep- antinociception. tive effects. Yohimbine precipitated withdrawal hyperalgesia in Multiple receptors involved in ␮,A1, and ␣2 tolerance clonidine-tolerant paws and withdrawal In the present study, we found that after induction of tolerance A similar proﬁle of receptor interactions was seen for tolerance with three hourly injections of clonidine (100 ng), administration and withdrawal to ␮,A1, and ␣2 antinociception. This was dem- of its receptor antagonist yohimbine (100 ng) precipitated a with- onstrated by examining which receptor agonists (␮,A1, and ␣2) drawal hyperalgesia, revealing the development of dependence could block antagonist-induced withdrawal hyperalgesia. (Fig. 2A). Yohimbine-induced withdrawal, in clonidine-tolerant paws, was blocked by coadministration of clonidine with yohimbine, as well Bidirectional cross-withdrawal for ␮,A1, and ␣2 as by coadministration of DAMGO and CPA with yohimbine tolerance/antinociception (Fig. 5A). Naloxone-induced withdrawal, in DAMGO-tolerant In paws made tolerant to clonidine, administration of the ␮- and paws, was blocked by coinjection of DAMGO with naloxone or A1-antagonists, naloxone and PACPX, respectively, precipitated clonidine with naloxone, but was not blocked by coadministration withdrawal hyperalgesia. In paws made tolerant to DAMGO and of CPA with naloxone (Fig. 5B). Similarly, PACPX-induced with- CPA, yohimbine precipitated withdrawal hyperalgesia (Fig. 2B). drawal in CPA-tolerant paws was blocked by coinjection of CPA These observations suggest that there is bidirectional cross- with PACPX or clonidine with PACPX, but was not blocked by ␮ ␣ withdrawal between ,A1, and 2 after the development of coadministration of DAMGO with PACPX (Fig. 5C). These data peripheral tolerance to their peripheral antinociceptive effects. suggest, as in previous experiments, that ␮, ␣2, and A1 receptors

Multiple receptors involved in ␮,A1, and are involved in ␣2 antinociception, tolerance, and withdrawal; that ␣2 antinociception ␮ and ␣2 receptors are involved in ␮ antinociception, tolerance,

Naloxone, yohimbine, and PACPX dose-dependently blocked the and withdrawal; and that A1 and ␣2 are involved in A1 antinoci- antinociceptive effects of their homologous agonists DAMGO, ception, tolerance, and withdrawal. clonidine, and CPA, respectively (Fig. 3A–C). However, in addition, naloxone (200 ng) also blocked clonidine antinociception but Antisense ODN treatment supports the hypothesis not CPA antinociception (Fig. 4A,C), and PACPX (100 ng) that multiple receptors are involved in ␮, ␣2, and blocked clonidine antinociception but not DAMGO antinocicep- A1 antinociception tion (Fig. 4A,B). Yohimbine (100 ng) blocked clonidine, Antisense ␮-opioid receptor ODN signiﬁcantly attenuated not

DAMGO, and CPA antinociception (Fig. 4A–C). These data only the ␮ antinociception but also ␣2 antinociception 24 hr after 740 J. Neurosci., January 15, 1997, 17(2):735–744 Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia

Figure 2. A, Yohimbine precipitates withdrawal hyperalgesia in clonidine tolerant paws. Effect of PGE2 (E2), clonidine plus PGE2 (ClϩE2), clonidine once hourly for 3 hr (Clx3), clonidine once hourly for 3 hr (Clx3), clonidine once hourly for 3 hr, and at the fourth hour yohimbine (Clx3, Yo)on mechanical paw-withdrawal threshold in the rat. B, Bidirectional cross-withdrawal develops among A1, ␣2, and ␮ antinociception. Effect of clonidine once hourly for 3 hr and at the fourth hour naloxone (Clx3, N), clonidine once hourly for 3 hr and at the fourth hour PACPX (Clx3, PACPX), clonidine once hourly for 3 hr (Clx3), DAMGO once hourly for 3 hr and at the fourth hour yohimbine (Dx3, Yo), DAMGO once hourly for 3 hr (Dx3), CPA once hourly for3hr(CPAx3), CPA once hourly for 3 hr and at the fourth hour yohimbine (CPAx3, Yo), and CPA once hourly for 3 hr (CPAx3) on mechanical paw withdrawal threshold in the rat.

Figure 3. ␮, ␣2, and A1 antagonists dose-dependently block ␮, ␣2, and A1 antinociception, respectively. A, Naloxone dose-dependently blocks DAMGO antinociception. Effect of PGE2 (E2), DAMGO plus PGE2 (DϩE2) and various doses of naloxone (1 ng to 1 ␮g), and DAMGO plus PGE2 (NϩDϩE2), on mechanical paw withdrawal threshold in the rat. B, Yohimbine dose-dependently blocks clonidine antinociception. Effect of PGE2 (E2), clonidine plus PGE2 (ClϩE2), and various doses of yohimbine (1 ng to 1 ␮g) and clonidine plus PGE2 (YoϩClϩE2) on mechanical paw withdrawal threshold in the rat. C, PACPX dose-dependently blocks CPA antinociception. Effect of PGE2 (E2), DAMGO plus PGE2 (CPAϩE2), and various doses of PACPX (1 ng to 1 ␮g) and CPA plus PGE2 (PACPXϩCPAϩE2) on mechanical paw withdrawal threshold in the rat. Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia J. Neurosci., January 15, 1997, 17(2):735–744 741

Figure 4. Multiple receptors are involved in ␮, ␣2, and A1 antinociception. A, Clonidine ␣2 antinociception is blocked not only by yohimbine but also by naloxone and PACPX. Effect of PGE2 (E2), clonidine plus PGE2 (ClϩE2), yohimbine plus clonidine plus PGE2 (YoϩClϩE2), naloxone plus clonidine plus PGE2 (NϩClϩE2), and PACPX plus clonidine plus PGE2 (PACPXϩClϩE2) on mechanical paw withdrawal threshold in the rat. B, DAMGO ␮ antinociception is blocked not only by naloxone but also by yohimbine. Effect of PGE2 (E2), DAMGO plus PGE2 (DϩE2), naloxone plus DAMGO plus PGE2 (NϩDϩE2), yohimbine plus DAMGO plus PGE2 (YoϩDϩE2), and PACPX plus DAMGO plus PGE2 (PACPXϩDϩE2) on mechanical paw withdrawal threshold in the rat. C, CPA A1 antinociception is blocked not only by PACPX but also by yohimbine. Effect of PGE2 (E2), CPA plus PGE2 (CPAϩE2), PACPX plus CPA plus PGE2 (PACPXϩCPAϩE2), naloxone plus CPA plus PGE2 (NϩCPAϩE2), and yohimbine plus CPA plus PGE2 (YoϩCPAϩE2) on mechanical paw withdrawal threshold in the rat.

Figure 5. Multiple receptors are involved in ␣2, ␮, and A1 tolerance and withdrawal. A, Yohimbine withdrawal is blocked not only by clonidine but also by DAMGO and CPA. Effect of clonidine once hourly for 3 hr and at the fourth hour yohimbine (Clx3,Yo), clonidine once hourly for 3 hr and at the fourth hour clonidine plus yohimbine (Clx3,ClϩYo), clonidine once hourly for 3 hr and at the fourth hour DAMGO plus yohimbine (Clx3,DϩYo), and clonidine once hourly for 3 hr and at the fourth hour CPA plus yohimbine (Clx3,CPAϩYo) on mechanical paw withdrawal threshold in the rat. B, Naloxone withdrawal is blocked not only by DAMGO but also by clonidine. Effect of DAMGO once hourly for 3 hr and at the fourth hour naloxone (Dx3,N), DAMGO once hourly for 3 hr and at the fourth hour DAMGO plus naloxone (Dx3,DϩN), DAMGO once hourly for 3 hr and at the fourth hour clonidine plus naloxone (Dx3,ClϩN), and DAMGO once hourly for 3 hr and at the fourth hour CPA plus naloxone (Dx3,CPAϩN) on mechanical paw withdrawal threshold in the rat. C, PACPX withdrawal is blocked not only by CPA but also by clonidine. Effect of CPA once hourly for 3 hr and at the fourth hour PACPX (CPAx3,PACPX), CPA once hourly for 3 hr and at the fourth hour CPA plus PACPX (CPAx3,CPAϩPACPX), CPA once hourly for 3 hr and at the fourth hour clonidine plus PACPX (CPAx3,ClϩPACPX), and CPA once hourly for 3 hr and at the fourth hour DAMGO plus naloxone (CPAx3,DϩPACPX) on mechanical paw withdrawal threshold in the rat. 742 J. Neurosci., January 15, 1997, 17(2):735–744 Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia

Figure 6. Antisense ␮ ODN treatment blocks not only ␮ antinociception but also ␣2 antinociception. A1 antinociception is unaffected. A, Effect of PGE2 (E2), DAMGO plus PGE2 (DϩE2), ␮-antisense (AS) ODN 1 ␮g intrathecally on alternate days ϫ 3 and DAMGO plus PGE2 [␮-(AS)x3,DϩE2], ␮-sense (S) ODN 1 ␮g intrathecally on alternate days ϫ 3, and DAMGO plus PGE2 [␮-(S)x3,DϩE2] on mechanical paw-withdrawal threshold. B, Effect of PGE2 (E2), clonidine plus PGE2 (ClϩE2), ␮-(AS) ODN 1 ␮g intrathecally on alternate days ϫ 3, and clonidine plus PGE2 [␮-(AS)x3,ClϩE2], ␮-(S) ODN 1 ␮g intrathecally on alternate days ϫ 3, and clonidine plus PGE2 [␮-(S)x3,ClϩE2] on mechanical paw-withdrawal threshold. C, Effect of PGE2 (E2), CPA plus PGE2 (CPAϩE2), ␮-(AS) ODN 1 ␮g intrathecally on alternate days ϫ 3, CPA plus PGE2 [␮-(AS)x3,CPAϩE2], ␮-(S) ODN 1 ␮g intrathecally on alternate days ϫ 3, and CPA plus PGE2 [␮-(S)x3,CPAϩE2] on mechanical paw withdrawal threshold in the rat.

the last injection of the antisense ODN. In contrast, A1 antinoci- without effect on PGE2 hyperalgesia, ␮ antinociception, or ␣2 ception was unaffected by ␮ ODN treatment (Fig. 6A–C). Sense antinociception.

␮-opioid receptor ODN was without effect on PGE2 hyperalgesia, The data from these antisense ODN experiments suggest that ␮ antinociception, or ␣2 antinociception (Fig. 6A–C). ␣2-adrenergic receptors are required for ␮ and A1 antinociception. Antisense ODN for the ␣2C-adrenergic receptor signiﬁcantly In addition, they suggest that the ␮-opioid receptor is required for ␣2 attenuated not only ␣2 antinociception but also ␮ and A1 antino- antinociception. Thus, multiple receptors are involved in the produc- ciception (Fig. 7A–C). Sense ␣2C-adrenergic receptor ODN was tion of antinociception by ␮, ␣2, and A1 agonists.

Figure 7. Antisense ␣2C ODN treatment blocks not only ␣2 antinociception but also ␮ and A1 antinociception. A, Effect of PGE2 (E2), clonidine plus PGE2 (ClϩE2), ␣2-(AS) ODN 1 ␮g intrathecally on alternate days ϫ 3, and clonidine plus PGE2 [␣2-(AS)x3,CClϩE2], ␣2-(S) ODN 1 ␮g intrathecally on alternate days ϫ 3, and clonidine plus PGE2 [␣2-(S)x3,ClϩE2] on mechanical paw withdrawal threshold in the rat. B, Effect of PGE2 (E2), DAMGO plus PGE2 (DϩE2), ␣2-(AS) ODN 1 ␮g intrathecally on alternate days ϫ 3, and DAMGO plus PGE2 [␣2-(AS)x3,DϩE2], ␣2-(S) ODN 1 ␮g intrathecally on alternate days ϫ 3, and DAMGO plus PGE2 [␣2-(S)x3,DϩE2] on mechanical paw withdrawal threshold in the rat. C, Effect of PGE2 (E2), CPA plus PGE2 (CPA ϩ E2), ␣2-(AS) ODN 1 ␮g intrathecally on alternate days ϫ 3, and CPA plus PGE2 [␣2-(AS)x3,CPAϩE2], ␣2-(S) ODN 1 ␮g intrathecally on alternate days ϫ 3, and CPA plus PGE2 [␣2-(S)x3,CPAϩE2] on mechanical paw withdrawal threshold in the rat. Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia J. Neurosci., January 15, 1997, 17(2):735–744 743

Figure 8. Schematic diagram of hypothesized topological/physical arrangement of the three receptors for peripheral antinociception in the cell membrane. ␮ (DAMGO),

␣2C (Clonidine), and A1 (CPA) agonism all result in peripheral antinociception mediated through a common second messenger pathway, leading to complete symmetrical cross-tolerance and cross-dependence. However, the asymmetrical interactions are proposed to be a result of the central posi-

tion of the ␣2C receptor leading to bidirectional interactions between this receptor and the two other receptors but no interac-

tion between the ␮ and A1 receptors.

DISCUSSION from neuronal membranes by morphine or naloxone In this study, we tested the hypothesis that peripheral antinoci- (Golombiowska-Nikitkin et al., 1980). Second, there may be phys- ception produced by ␮, ␣2, and A1 agonists exhibit cross-tolerance ical interactions between the receptors in the cell membrane. Such after repeated exposure to these agents. This hypothesis arose interactions have been suggested to explain effects of agonist from previous data demonstrating that ␮, ␣2, and A1 receptors combinations that are greater than additive (synergistic) or less can signal via a common second messenger, namely activation of than additive (antagonistic) than the effects seen at the different an inhibitory G-protein (Sharma et al., 1975; Law et al., 1981; receptors. For example, an ␣2 agonist attenuates both A1 and ␮ Childers and Rivere, 1989; Mankmann et al., 1988). In all exper- mediated inhibition of norepinephrine release from sympathetic iments evaluating cross-tolerance and cross-dependence, com- nerve endings; this interaction has been suggested to occur at the plete symmetry was found for ␮, ␣2, and A1. These findings level of the receptor (Bucher et al., 1992). Furthermore, Bentley support the hypothesis that there is a common signaling pathway et al. (1983) have hypothesized that ␣-adrenoceptor and opioid for these three receptors. Although the mechanisms of tolerance receptors may be linked, either via second messenger systems or and dependence in primary afferent nociceptors is unknown, in physically in the membrane. other systems the protein kinase C second messenger system has We hypothesize from the data in our experiments that the ␣2 been implicated (Mao et al., 1995; Mayer et al., 1995). The role of receptor is arranged topologically between the ␮ and A1 re- this second messenger system in the tolerance and dependence to ceptors to form a receptor complex (Fig. 8). This hypothesis is peripheral antinociception is currently being investigated. These supported by the observation that antisense oligodeoxynucle- data also suggest that clinically these pharmacologies may not be otides against ␮-opioid receptor reduced not only ␮ antinoci- cross-substituted in dependent individuals. ception, but also ␣ antinociception, while preserving A an- As a control for the cross-withdrawal experiments, we tested 2 1 tinociception, whereas the ␣ antisense oligodeoxynucleotide whether there was blockade of antinociception by heterologous 2C reduced the antinociceptive effects of all three receptor sys- antagonists in naive animals. Unexpectedly, we found that the ␣2 tems, ␣2,A1, and ␮. The preservation of A1 antinociception antagonist blocked not only ␣2 antinociception but also A1 and ␮ after ␮ antisense treatment but not after ␣2C antisense treat- antinociception, that both ␮ and A1 antagonists blocked ␣2 antinociception, and that there was no such heterologous antago- ment suggests that the effect of receptor attenuation in the terminal is of short-range topologically. The same profile of nism between ␮ and A1 ligands in antinociception. The absence of interactions between ␣2, ␮, and A1 were found in assessing the an interaction between ␮ and A1 antinociception is unlikely to be a result of an inadequate dose of antagonist because even at a very ability of a heterologous agonist to block withdrawal induced high dose (1 ␮g) no cross-antagonism was observed (unpublished by the homologous antagonist in a tolerant paw. observations). In summary, we have demonstrated that there is a symmetry for There are known mechanisms by which antagonists can heter- the three agonists for production of tolerance and dependence. ologously antagonize the actions of an agonist at another receptor There was also an unexpected interaction, which seems to occur at class. First, when a receptor ligand is not highly selective, at the level of the receptor, between ␣2 and ␮ receptors and ␣2 and sufficiently high doses it will bind to another receptor to displace A1 receptors, but not between ␮ and A1 receptors. The results a heterologous ligand (Cicero et al., 1974; Spiehler et al., 1978; suggest the hypothesis that these three receptors may be coupled Blank et al., 1983). However, clonidine binding is not displaced physically in the plasma membrane. 744 J. Neurosci., January 15, 1997, 17(2):735–744 Aley and Levine • Multiple Receptor Interactions in Peripheral Analgesia

REFERENCES Law PY, Wu J, Koehler E, Loh HH (1981) Demonstration and charac- Aley KO, Green PG, Levine JD (1995) Opioid and adenosine peripheral terization of opiate inhibition of the striatal adenylate cyclase activity. antinociception are subject to tolerance and withdrawal. J Neurosci J Neurochem 36:1834–1846. 15:8031–8038. Lingen LU, Ordway G (1995) Antisense knock-down of ␣2C- Bentley GA, Newton SH, Starr J (1983) Studies on the antinociceptive adrenoceptors in rat striatum. Neurosci Abstr 21:1612. action of ␣-agonist drugs and their interactions with opioid mechanisms. Mankman MH, Dvorkin B, Crain SM (1988) Modulation of adenylate Br J Pharmacol 79:125–134. cyclase activity of mouse spinal cord-ganglion explants by opioids, Blank MS, Diez JA, Roberts DL (1983) Monoaminergic antagonists serotonin and pertussis toxin. Brain Res 445:303–313. which block naloxone-induced release of luteinizing hormone bind Mao J, Price DD, Phillips L, Lu J, Mayer D (1995) Increase in protein selectively to hypothalamic opiate receptors. Brain Res 279:153–158. kinase C ␥ immunoreactivity in the spinal cord of rats associated with Bucher B, Corriu C, Stoclet JC (1992) Prejunctional opioid ␮-receptors tolerance to analgesic effects of morphine. Brain Res 677:257–267. and adenosine A1-receptors on the sympathetic nerve endings of the rat Mayer DJ, Mao J, Price DD (1995) The development of morphine tol- tail artery interact with the ␣2-adrenoceptors. Naunyn Schmiedebergs erance and dependence is associated with translocation of protein Arch Pharmacol 345:37–43. kinase C. Pain 61:365–374. Childers SR, LaRiviere G (1984) Modiﬁcation of guanine nucleotide- Pitchford S, Levine JD (1991) Prostaglandins sensitize nociceptors in cell regulatory components in brain membranes. I. Relationship of culture. Neurosci Lett 132:105–108. guanosine 5Ј-triphosphate effects on opiate receptor binding and cou- Rossi G, Pan YX, Cheng J, Pasternak GW (1994) Blockade of morphine pling receptors with adenylate cyclase. J Neurosci 4:2764–2771. analgesia by an antisense oligodeoxynucleotide against the ␮ receptor. Cicero TJ, Wilcox CE, Meyer ER (1974) Effect of ␣-adrenergic blockers Life Sci 54:375–379. on naloxone-binding in brain. Biochem Pharmacol 23:2349–2352. Sharma SK, Nirenberg M, Klee WA (1975) Morphine receptors as reg- Gold MS, White DM, Ahlgren SC, Guo M, Levine JD (1994) ulators of adenylate cyclase activity. Proc Natl Acad Sci USA Catecholamine-induced mechanical sensitization of cutaneous nocicep- 72:590–594. tors in the rat. Neurosci Lett 175:166–170. Spiehler V, Fairhurst AS, Randall LO (1978) The interaction of phe- Golombiowska-Nikitkin K, Pilc A, Ventulani J (1980) Opiate and speciﬁc noxybenzamine with the mouse brain opiate receptor. Mol Pharmacol receptor binding of [3H]-clonidine. J Pharm Pharmacol 32:70–71. 14:587–595. Khasar SG, Green PG, Chou B, Levine JD (1995) Peripheral nociceptive Taiwo YO, Coderre TJ, Levine JD (1989) The contribution of training to effects of ␣2-adrenergic receptor agonists in the rat. Neuroscience sensitivity in the nociceptive paw-withdrawal test. Brain Res 66:427–432. 487:148–151. Khasar SG, Gold MS, Dastmalchi S, Levine JD (1996) Selective attenu- Taiwo YO, Levine JD (1990) Direct cutaneous hyperalgesia induced by ation of ␮-opioid receptor-mediated effects in rat sensory neurons by adenosine. Neuroscience 38:757–762. intrathecal administration of antisense oligodeoxynucleotides. Neurosci Yaksh TL, Rudy TA (1976) Chronic catheterization of the spinal sub- Lett 218:17–20. arachnoid space. Physiol Behav 17:1031–1036.