Clonidine and Dexmedetomidine Produce Antinociceptive Synergy in Mouse Spinal Cord Carolyn A

PAIN MEDICINE

Anesthesiology 2009; 110:638–47 Copyright © 2009, the American Society of Anesthesiologists, Inc. Lippincott Williams & Wilkins, Inc. Clonidine and Dexmedetomidine Produce Antinociceptive Synergy in Mouse Spinal Cord Carolyn A. Fairbanks, Ph.D.,* Kelley F. Kitto,† H. Oanh Nguyen, B.S.,‡ Laura S. Stone, Ph.D.,§ George L. Wilcox, Ph.D.

Background: Synergy between drugs manifests with in- disparate anatomical sites or distinct receptors coresid- creased potency and/or efficacy of the combination relative to ing at a common anatomical location. Examples of well- either agonist given alone. Synergy is typically observed between drugs of different classes, as is the case with the ␣-adren- described synergistic agonist pairs include selective ago- ergic–opioid receptor synergy often observed in preclinical nists of the ␮- and ␦-opioid receptor subtypes as well as Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 studies. However, rare studies report synergy between agonists either of those subtypes combined with agonists target- of the same class. The current study examined the analgesic ␣ ing the 2-adrenergic receptors (ARs). interaction between two intrathecally injected ␣ -adrenergic 2 The analgesic and anesthetic properties of ␣ AR-selec- receptor (AR) agonists previously thought to act at the same 2 receptor subtype when given spinally. tive agonists have been known for decades. Develop- Methods: Mice were given clonidine, dexmedetomidine, or ment of clinical applications of these agonists remains an the combination spinally to evaluate the interaction between area of interest, particularly as adjuvants for pain man- 1 these two agonists. The ED50 values were calculated, and the agement and as anesthetic-sparing agents. In contrast to interactions were tested by isobolographic analysis. The ro- the opioid receptor–selective agonists, definition of the tarod test was performed in the same mice after the completion ␣ of analgesic assessment to assess motor or sedative effects. pharmacologic profile of each 2AR agonist has been These experiments were performed in outbred mice as well as limited because of poor ligand selectivity across the ␣ ␣ ␣ ␣ ␣ ␣ 2 in mice with mutant 2AARs, 2CAR knockout mice, or wild-type three 2AR subtypes: 2AAR, 2BAR, and 2CAR. The controls. Finally, analgesic cross-tolerance between clonidine ␣ 2AR subtypes are differentially expressed in specific and dexmedetomidine was evaluated. regions of the central nervous system. For example, in Results: Clonidine and dexmedetomidine interacted syner- ␣ the spinal cord, ␣ ARs seem to be principally of pri- gistically in all lines except the 2CAR knockout line, implicat- 2A ␣ ␣ ing 2CARs in the interaction. In addition, clonidine and dexme- mary afferent neuron origin, whereas 2CARs seem to be detomidine did not show analgesic cross-tolerance in the expressed primarily on neurons intrinsic to the spinal outbred strain, suggesting that the two drugs have distinct cord.3 The evidence for ␣ AR expression in spinal cord mechanisms of action. 2B Conclusions: The current study introduces a new synergistic nerve terminals and intrinsic spinal neurons is not con- ␣ 4,5 ␣ 6 agonist pair, clonidine–dexmedetomidine. These two drugs clusive. Activation of both 2AARs and 2CARs has ␣ seem to require the 2AAR for spinal analgesia when given been reported to result in antinociception. Therefore, it separately; when delivered as a combination, the resultant syn- is reasonable to propose that concurrent participation of ␣ ergistic interaction requires the 2CAR as well. ␣ ␣ 2AARs and 2CARs could result in analgesic synergy. ␣ 4,5 Support for a positive interaction between 2AARs and SYNERGISTIC drug interactions result in enhanced po- ␣ 2CARs is provided in a previous report that evaluated tency and/or efficacy when one agent is given together ␣ 7 interactions between two 2-adrenergic agonists that with another. Therapeutic application of synergistic ␣ were thought to act at different 2AR subtypes based on combinations carries the expectation of efficacy at re- differences in the pharmacology of their antagonist sen- duced doses and, theoretically, reduced side effects. sitivity. To approach this question systematically, we Although the mechanisms underlying synergistic inter- ␣ have initiated a broad evaluation of several 2AR agonist actions are not well understood, synergy is thought to ␣ ␣ combinations in mouse lines deficient in 2AAR or 2CAR result from simultaneous action of the two agents at two function. As part of this larger program, the current distinct sites, such as a common receptor located at study evaluated the interaction between intrathecally administered clonidine and dexmedetomidine. Previous * Associate Professor, † Assistant Scientist, ‡ Research Coordinator, Depart- ␣ studies of 2AAR mutant mice have been interpreted to ments of Pharmaceutics, Pharmacology, and Neuroscience, Professor, Depart- ments of Neuroscience, Pharmacology, and Dermatology, University of Minne- indicate that the potency and/or efficacy of both of these sota. § Assistant Professor, Alan Edwards Centre for Research on Pain and ␣ agonists is primarily dependent on 2AAR activation, Department of Pharmacology, McGill University, Montreal, Quebec, Canada. particularly when administered intrathecally. Because of Received from the Departments of Pharmaceutics, Pharmacology, and Neuro- science, University of Minnesota, Minneapolis, Minnesota. Submitted for publi- this prevailing view, we did not expect that coadminis- cation April 9, 2008. Accepted for publication August 19, 2008. Supported by tration of clonidine with dexmedetomidine would result grant Nos. R01-DA-15438 (Dr. Wilcox) and K01-DA-00509 to (Dr. Fairbanks) from the National Institute on Drug Abuse, Bethesda, Maryland. Zeneca Pharmaceuti- in a synergistic analgesic interaction. Our observations cals, Wilmington, Delaware, provided dexmedetomidine. indicate, however, that this combination produces de- Address correspondence to Dr. Fairbanks: Department of Pharmaceutics, finitive and replicable synergistic analgesia in several 9-177 Weaver Densford Hall, 308 Harvard Street Southeast, Minneapolis, Minne- sota 55455-0217. [email protected]. Information on purchasing reprints may be separate strains of mice: CD-1 Institute of Cancer Re- found at www.anesthesiology.org or on the masthead page at the beginning of ␣ search (ICR) outbred mice, mice deficient in the 2AAR this issue. ANESTHESIOLOGY’s articles are made freely accessible to all readers, for ␣ personal use only, 6 months from the cover date of the issue. or the 2CAR subtype, and their wild-type (WT) controls.

Anesthesiology, V 110, No 3, Mar 2009 638 CLONIDINE–DEXMEDETOMIDINE SYNERGY IN MICE 639

Further, the potential for cross-tolerance between the ago- Nociceptive Assay nists was assessed after chronic intrathecal delivery of ei- Nociceptive responsiveness was tested in the SP noci- ther agonist. Finally, the interaction between clonidine and ceptive test. The SP assay is a sensitive indicator of dexmedetomidine on a measure of sedation and motor milder analgesics.12 SP (10–20 ng) was injected intra- coordination (accelerating rotarod) was also evaluated. thecally to produce approximately 40–60 behaviors (scratches and bites directed to the hindquarters) in the ﬁrst minute after injection. The dose of SP required to Methods and Materials produce this number of behaviors was conﬁrmed with Animals each new experiment. Coadministration of opioid or Experimental subjects were 20- to 25-g male ICR mice adrenergic analgesics dose-dependently inhibits those 13

(Harlan, Madison, WI) or 15- to 20-g male and female mice behaviors. To test the ability of dexmedetomidine and Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 (sex matched) with either a mixed C57BL/6-129/Sv genetic clonidine to inhibit SP-induced behavior, the drugs were ␣ ␣ background ( 2AAR-WT or 2AAR-D79N) or a pure coadministered with SP and inhibition was expressed as ␣ ␣ C57BL/6 background ( 2CAR-WT or 2CAR-knockout a percent of the mean response of the control group [KO]). Animals were maintained on a 12-h light/dark cycle (determined with each new experiment) according to ␣ and had unlimited access to food and water. The 2AAR- the following equation: D79N mutant mice had been generated by hit-and-run gene % Inhibition targeting as previously described8 on a hybrid C57BL/6- 129/Sv background. WT animals of the same mixed back- ϭ ͓͑Control Ϫ Experimental͒ ⁄ Control͔ ϫ 100. ␣ ␣ ground were used as controls ( 2AAR-WT). The 2CAR-KO mice were developed at Stanford University, Palo Alto, Sedation/Motor Impairment Assay California,9 and were purchased from Jackson Labs, Bar In the same mice that received SP stimulation, doses of Harbor, Maine, after 17 generations of backcrossing to clonidine, dexmedetomidine, and their combination C57BL/6 background. C57BL/6 mice pair-bred within our were tested for impairment of rotarod performance. In ␣ facility were used as WT controls ( 2CAR-WT). Breeding such experiments, the animals were trained the day pairs were established, and pups were weaned between 2 before experimentation to walk 300 s on the accelerat- and 3 weeks of age. Within each experiment, animals were ing rotarod, typically requiring three trials to learn the age- and sex-matched across groups. Animals were used no behavior. The following day, the drugs were adminis- more than twice. In each case, a rest period of at least 1 tered with SP. After completion of the 1-min SP-evoked week was used, and the animals were randomized across scratching and biting analysis, the mice were run on the treatment groups. Although the use of transgenic or KO rotarod test. mice may result in compensatory changes, we chose to use Motor impairment or sedation was expressed as inhi- these mouse lines because we have extensively character- bition of the subjects’ ability to remain on the acceler- ized their spinal neuropharmacology,4–6 and they have ating rotarod; baseline latencies to fall were typically at ␣ been widely used by other groups with interest in 2AR- or near the cutoff of 300 s. Percent inhibition was ex- mediated antinociception and antihypertensive effects (for pressed as a percent of the baseline latency of each review, see Kable et al.10). Therefore, the results presented mouse (determined prior to each new experiment) ac- in this study are directly comparable to the previous liter- cording to the following equation: ature. These experiments were approved by the Institu- tional Animal Care and Use Committee of the University of % Inhibition Minnesota, Minneapolis, Minnesota. Subjects were housed ϭ ͓͑Baseline Ϫ Experimental͒ ⁄ Baseline͔ ϫ 100. in groups of four in 25 ϫ 48 ϫ 15-cm plastic cages in a temperature- and humidity-controlled environment, were Dose–Response Analysis maintained on a 12-h light/dark cycle, and had free access Individual dose and/or time points are expressed as to food and water. mean and SEM. ED50 values and confidence limits were calculated according to the graded dose–response Chemicals method of Tallarida and Murray14 on the linear portion Clonidine HCl (2-[2,6-dichloroaniline]-2-imidazoline) of each dose–response curve. Statistical comparisons of and substance P (SP) were purchased from Sigma Chem- potencies are based on the confidence limits of the ED50 ical Co. (St. Louis, MO). SP was dissolved in acidified values. A minimum of three doses were used for each saline. Zeneca (Wilmington, DE) donated the dexme- drug or combination of drugs. A minimum of 50% was detomidine [1-(S)-4-[1-(2,3-dimethylphenyl) ethyl]1H- set for a drug to be classified as efficacious. imidazole]. Clonidine and dexmedetomidine were dissolved in 0.9% saline. All drugs were administered intra- Isobolographic Analysis thecally by direct lumbar puncture in a 5-␮l volume in Dose–response curves were constructed for each ago- 11 conscious mice. nist administered alone; the ED50 values were calculated

Anesthesiology, V 110, No 3, Mar 2009 640 FAIRBANKS ET AL.

tions14,15) necessitates this manipulation. When testing an interaction between two drugs, a theoretical additive

ED50 value 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 with the observed experimental ED50 value of the combination. These values are based on the total dose of both drugs. An interaction is considered synergistic if the Ͻ experimental ED50 is signiﬁcantly less (P 0.05) than the calculated theoretical additive ED50 value.

Fig. 1. Clonidine (Clon) and dexmedetomidine (Dex) interact Visualization of drug interactions can be facilitated and Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 synergistically when given spinally to Institute of Cancer Re- enhanced by graphical representation of isobolographic search (ICR) mice. (A) Clonidine () and dexmedetomidine (f) analysis. This representation depicts the ED of each inhibited substance P behavior in a dose-dependent manner. 50 The agonists were then coadministered at a constant clonidine: agent as the x- or y-intercept. For example, figure 1B ؉ □ dexmedetomidine dose ratio of 1:1 [ Clon ( Dex)] based on presents the ED50 of clonidine as the y-intercept and the the potency ratio between agonists. Note that the combination ED of dexmedetomidine as the x-intercept. The line dose–response curves are plotted as the doses of clonidine used 50 in the presence of dexmedetomidine. The corresponding Dex connecting these two points depicts the dose combina- ؉ Clon) curve is identical and not shown. (B) Isobolographic tions expected to yield 50% efficacy if the interaction is) analysis applied to the data from A. The y-intercept represents purely additive and is called the theoretical additive the ED50 for clonidine, and the x-intercept represents the ED50 line. The theoretical additive ED and its confidence for dexmedetomidine. The observed combination ED50 ( ) was 50 significantly lower (P < 0.05, t test) than the theoretical additive interval are determined mathematically and plotted span- Œ ED50 ( ), indicating that the interaction is synergistic in ICR ning this line. The observed ED50 for the combination is mice. See table 1 for ED50 values. Group sizes ranged from five intrathecal. plotted at the corresponding x,y coordinates along with ؍ .to eight mice. i.t its 95% confidence interval for comparison to the theo- and used to determine the potency ratio between the retical additive ED50. All dose–response and isobolo- agonists (e.g., fig. 1A). This ratio was then maintained graphic analyses were performed with the FlashCalc when both agonists were administered in combination, a 4.5.3 pharmacologic statistics software package16,17 gen- third dose–response curve was constructed, and an ex- erously supplied by Michael Ossipov, Ph.D. (Professor, perimentally derived combination ED50 was calculated. University of Arizona, Tucson, Arizona). To test for interactions between agonists, the ED50 values and SE for all dose–response curves were arithmet- Chronic Clonidine or Dexmedetomidine ically arranged around the ED50 value using the equation Tolerance Induction 10 ϫ ϫ 15 (ln ED50) (SE of log ED50). Isobolographic anal- To induce spinal clonidine or dexmedetomidine toler- ysis (the standard for the evaluation of drug interac- ance, clonidine or dexmedetomidine (10 nmol in 5 ␮l)

Table 1. Summary of Clonidine–Dexmedetomidine Antinociceptive Interactions

Probe Drug, Intrathecal, nmol ED50 Clonidine (95% CI) ED50 Dexmedetomidine (95% CI) Interaction

Figure 1, ICR mice Single drug 3.0 (2.1–3.9) 3.8 (1.3–6.3) Clonidine–dexmedetomidine, 1:1 ratio Synergistic Observed combination 0.0045 (0.0001–0.0183)* Theoretical additive 1.7 (1.3–2.1) ␣ Figure 2, 2AAR-WT Single drug 1.2 (0.45–1.9) 1.0 (0.43–1.7) Clonidine–dexmedetomidine, 1:1 ratio Synergistic Observed combination 0.18 (0.1–0.26)* Theoretical additive 0.54 (0.34–0.74) ␣ Figures 4A and B, 2CAR-WT Single drug 1.7 (1.3–2.1) 1.9 (1.3–2.5) Clonidine–dexmedetomidine, 1:1 ratio Synergistic Observed combination 0.16 (0.11–0.23)* Theoretical additive 0.90 (0.73–1.1) ␣ Figures 4C and D, 2CAR-KO Single drug 5.3 (4.1–6.5) 4.4 (3.5–5.3) Clonidine–dexmedetomidine, 1:1 ratio Subadditive Observed combination 3.6 (2.2–5.0)* Theoretical additive 1.71 (1.3–2.1)

* Signiﬁcant difference from theoretical additive by Student t test, P Ͻ 0.05. AR ϭ adrenergic receptor; CI ϭ conﬁdence interval; ICR ϭ Institute of Cancer Research; KO ϭ knockout; WT ϭ wild type.

Anesthesiology, V 110, No 3, Mar 2009 CLONIDINE–DEXMEDETOMIDINE SYNERGY IN MICE 641

Fig. 2. Clonidine (Clon) produces antinociceptive synergy with dexmedetomidine ␣ (Dex) in 2A-adrenergic receptor (AR) wild-type (WT) mice. (A) Clonidine () and dexmedetomidine (f) inhibited substance P behavior in a dose-dependent manner. The agonists were then coadministered at a constant clonidine: dexmedetomidine dose ratio of 1:1 [□ Clon (؉ Dex)] based on the potency ratio between agonists. Note that the combination dose–response curves are plotted as the doses of clonidine used in the presence of dexmedetomidine. The corresponding Dex (؉ Clon) curve is identical and not shown. (B) Isobolographic analysis applied to the data from A. The y-intercept Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 represents the ED50 for clonidine, and the x-intercept represents the ED50 for dexmedetomidine. The observed combination ED50 ( ) Œ was signiﬁcantly lower (P < 0.05, t test) than the theoretical additive ED50 ( ), indicating that the interaction is synergistic in ␣ 2AAR-WT mice. See table 1 for ED50 values. (C) Substance P–induced behavior was challenged by intrathecally (i.t.) administered ␣ f clonidine, dexmedetomidine, or both in 2AAR-D79N mice. Neither clonidine ( ) nor dexmedetomidine ( ) inhibited the behavior. .The coadministration of the agonists in a dose ratio of 1:1 (□ Clon ؉ Dex) did not produce appreciable inhibition of the behavior Group sizes ranged from ﬁve to eight mice.

␣ was delivered intrathecally once on experimental day 1 and 2AAR-D79N mice were (1) to determine whether and twice daily on experimental days 2 and 3. Repeated the synergistic interaction was observable across injections were separated by at least 8 h. A separate mouse strains and (2) to determine whether the com- group of mice received an equal number of injections of bination demonstrated any efficacy in mice lacking ␣ saline as a control group. On experimental day 4, full 2AAR. Because clonidine consistently demonstrates ␣ dose–response curves were constructed for each agonist no efficacy in 2AAR-D79N mice and dexmedetomi- in each pretreatment group. The antinociceptive poten- dine is only efficacious at high doses, the expectation cies (ED50 values) of clonidine and dexmedetomidine to was that the combination would not yield significant inhibit SP-evoked behaviors were compared between mice efficacy in those mice; nonetheless, it was important pretreated with saline or clonidine or dexmedetomidine. to test the possibility that the combination resulted in a significantly different pharmacologic profile than either agonist alone. We first determined the potency Results of each agonist to inhibit SP-evoked behavior in ␣ Clonidine–Dexmedetomidine Analgesic Synergy 2AAR-WT mice. As expected, clonidine and dexme- Clonidine Produces Analgesic Synergy with detomidine inhibited the behavior with comparable Dexmedetomidine in ICR Mice. We first determined potency and efficacy (fig. 2A). Coadministration of the potency of each agonist to inhibit SP-evoked behav- clonidine with dexmedetomidine resulted in combina- ior in ICR mice. As expected, clonidine and dexmedeto- tion dose–response curves shifted approximately sev- midine inhibited the behavior with comparable potency enfold to the left compared with each agonist given separately (fig. 2A and table 1). The isobologram in and efficacy (fig. 1A). The calculated ED50 values of these dose–response curves formed the basis for the equief- figure 2B illustrates that the ED50 value of the ob- fective dose ratios used in the respective combinations served combination differs significantly from the cal- (table 1). Coadministration of clonidine with dexmedeto- culated theoretical additive ED50 value, indicating a midine resulted in combination dose–response curves synergistic interaction (fig. 2B and table 1; Student t shifted approximately 700-fold to the left compared with test, P Ͻ 0.05). The synergistic interaction of the ␣ each agonist given separately (fig. 1 and table 1). The clonidine–dexmedetomidine combination in 2AAR-WT isobologram in figure 1B illustrates that the ED50 value of mice confirms that the observation was not unique to the observed combination differs significantly from the ICR mice. Although the magnitude of synergism is calculated theoretical additive ED50 value, indicating a syn- significantly different (100-fold) across these two ergistic interaction (fig. 1B and table 1; Student t test, P Ͻ strains, the observation of significant synergy for this 0.05). This experiment was replicated in a separate group combination is consistent. This difference also profiles of mice with comparable outcomes (synergism). The ro- the importance of evaluating combinations across mul- bust synergistic interaction of the clonidine–dexmedetomi- tiple strains. Consistent with our previous reports, dine combination suggests a second spinal site of action for neither clonidine nor dexmedetomidine demonstrates one of the two agonists. ␣ antinociceptive efficacy in the 2AAR-D79N mice Clonidine–Dexmedetomidine Coadministration when given either alone or as a 1:1 combination, even ␣ in 2AAR-WT Mice. The objectives for testing the at relatively high doses (10, 30, and 100 nmol, intra- ␣ clonidine–dexmedetomidine combination in 2AAR-WT thecally; fig. 2C).

Anesthesiology, V 110, No 3, Mar 2009 642 FAIRBANKS ET AL.

Fig. 3. Chronic intrathecal (i.t.) clonidine (Clon) or dexmedetomidine (Dex) does not evoke mutual cross-tolerance. (A) Clonidine intrathecal tolerance. The potency of clonidine was significantly reduced in mice pretreated (Prtx) with repeated injections of clonidine (Œ) relative to saline pretreatment (), indicating the development of analgesic tolerance. In contrast, the potency of dexmedetomidine in mice pretreated with repeated injections of clonidine (□) did not differ relative to mice pretreated with saline (f), confirming the lack development of analgesic cross-tolerance. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 (B) Dexmedetomidine intrathecal tolerance. The potency of dexmedetomidine was significantly reduced in mice pretreated with repeated injections of dexmedetomidine (□) relative to saline pretreatment (f), indicating the development of analgesic tolerance. In contrast, the potency of clonidine in mice pretreated with repeated injections of dexmedetomidine (Œ) did not differ relative to mice pretreated with saline (), confirming the lack development of analgesic cross-tolerance. The ED50 values for the dose–response groups are presented in table 2. Group sizes were eight mice per dose group.

Mechanism of Clonidine–Dexmedetomidine remained largely unchanged (fig. 3A and table 2). Simi- Analgesic Synergism larly, 3-day spinal pretreatment with dexmedetomidine Clonidine and Dexmedetomidine Do Not Evoke significantly reduced the potency of “probe” doses of Chronic Analgesic Cross-Tolerance. The observation dexmedetomidine (21-fold tolerance), but the analgesic of synergy between clonidine and dexmedetomidine response to clonidine was not significantly altered (fig. ␣ suggests that a receptor other than the 2AAR is involved 3B and table 2). This lack of cross-tolerance suggests in the interaction. In situations where two agonists act that, despite their apparent shared reliance on spinal ␣ primarily at the same receptor, chronic administration of 2AARs when given separately, clonidine-evoked or one agonist usually elicits cross-tolerance to the other.18 dexmedetomidine-evoked antinociception requires par- Conversely, in cases where two agonists act at different ticipation of a second distinct receptor. receptor sites, chronic exposure to one agonist typically Clonidine Produces Analgesic Synergy with ␣ does not invoke chronic tolerance to the other (e.g., Dexmedetomidine in C57Bl/6 but Not 2CAR-KO ␮-opioid receptor (MOP), ␦-opioid receptor (DOP)19,20), Mice. A logical candidate for the second receptor is the ␣ although minor cross-tolerance is sometimes observed, 2CAR, given its localization in spinal cord and previous ␣ perhaps because of changes in convergent downstream studies illustrating that 2CAR activation can result in ␣ 20–22 6,23 signaling pathways (e.g., MOP– 2AAR ). Therefore, antinociception. Therefore, we tested for clonidine– ␣ to evaluate whether clonidine and dexmedetomidine dexmedetomidine synergy in 2CAR-KO mice and their may act on the same or different receptors, we con- WT controls (C57BL/6 mice). The clonidine–dexme- ducted an evaluation of analgesic tolerance to clonidine detomidine combination demonstrated significant anal- ␣ or dexmedetomidine after repeated chronic exposure to gesic synergy in 2CAR-WT mice (figs. 4A and B and table ␣ spinally administered clonidine (fig. 3A) or dexmedeto- 1), as was the case in ICR and 2AAR-WT mice. In ␣ midine (fig. 3B) in ICR mice. Whereas 3-day spinal pre- contrast to the lack of efficacy observed in the 2AAR- treatment with clonidine significantly reduced the po- D79N mice (fig. 2C), the analgesic potency of clonidine tency of “probe” doses of clonidine (16-fold tolerance), and dexmedetomidine decreased only twofold to three- ␣ the analgesic dose–response curve for dexmedetomidine fold (though significantly) in 2CAR-KO mice relative to

Table 2. Summary of Clonidine and Dexmedetomidine Tolerance ED50 Values and Potency Shifts

Saline– Clonidine– Saline– Clonidine– Cross- Pretreatment Probe Clonidine Clonidine Tolerance Dexmedetomidine Dexmedetomidine Tolerance

Figure 3A, clonidine 0.93 (0.0–1.3) 16 (14–18)* 16-fold reduced 1.33 (1.0–1.8) 1.26 (0.91–1.73) No signiﬁcant tolerance potency potency shift

Saline– Dexmedetomidine– Saline– Dexmedetomidine– Cross- Pretreatment Probe Dexmedetomidine Dexmedetomidine Tolerance Clonidine Clonidine Tolerance

Figure 3B, dexmedetomidine 0.71 (0.54–0.93) 15 (13–17)* 21-fold reduced 0.98 (7.3–1.3) 1.5 (1.1–2.1) No signiﬁcant tolerance potency potency shift

* Signiﬁcant difference in relative potency.

Anesthesiology, V 110, No 3, Mar 2009 CLONIDINE–DEXMEDETOMIDINE SYNERGY IN MICE 643

(ﬁgs 4C and D and table 1). The potency of the clonidine–dexmedetomidine combination was not altered relative to that of either agonist given alone; the

combination ED50 value was signiﬁcantly higher than that of the theoretical additive ED50 value. This result suggests that the clonidine–dexmedetomidine synergis- ␣ tic interaction requires the presence of 2CARs and that ␣ in the absence of 2CARs the two drugs may act at the ␣ same receptor, presumably the 2AAR.

Clonidine–Dexmedetomidine Interactions in Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 Other Assays Clonidine–Dexmedetomidine Interactions in the Rotarod Assay of Sedation and Motor Impairment. ␣ In addition to their analgesic effects, 2AR agonists affect multiple physiologic systems, including the central nervous system (sedation, cardiovascular effects, addiction and withdrawal responses). In the current study, the rotarod test, which has been previously used as a measure of both sedation and motor impairment,24 was used to assess the sedative and/or motoric effects of the agonists or their combination immediately after SP nocicep- Fig. 4. Clonidine (Clon) and dexmedetomidine (Dex) interact ␣ tive testing. In outbred ICR mice (fig. 5A), clonidine and synergistically in 2C-adrenergic receptor (AR) wild-type (WT) ␣ dexmedetomidine each produced a mild reduction in but not 2CAR knockout (KO) mice. (A) Clonidine ( ) and dexmedetomidine (f) inhibited the substance P behavior in a rotarod performance at the highest dose tested (10 dose-dependent manner. The agonists were then coadminis- nmol); higher doses were not tested. The clonidine– tered at a constant clonidine:dexmedetomidine dose ratio of 1:1 -Clon (؉ Dex)] based on the potency ratio between agonists. dexmedetomidine combination reduced rotarod perfor □] Note that the combination dose–response curves are plotted as mance only 30% at the highest combination dose (1 the doses of clonidine used in the presence of dexmedetomi- nmol of each drug) tested, which produced approxi- (dine. The corresponding Dex (؉ Clon) curve is equivalent. (B Isobolographic analysis applied to the data from A. The y- mately 90% antinociception (fig. 1A); potentiation was intercept represents the ED50 for clonidine, and the x-intercept evident at 0.01 and 1 nmol. We distinguish this interac- represents the ED50 for dexmedetomidine. The observed com- tion in rotarod from the synergistic analgesic interaction bination ED50 ( ) was significantly lower (P < 0.05, t test) than Œ by referring to the former as potentiation. In ␣ AR-WT the theoretical additive ED50 ( ), indicating that the interaction 2A ␣ is synergistic in 2CAR-WT mice. (C) Substance P–induced be- mice (fig. 5B), clonidine reduced rotarod performance havior was challenged by intrathecally (i.t.) administered ␣ approximately 70% at the highest dose tested (10 nmol), clonidine, dexmedetomidine, or both in 2CAR-KO mice. Clonidine () and dexmedetomidine (f) inhibited the behavior whereas dexmedetomidine produced only partial reduc- in a dose-dependent manner. The agonists were then coadmin- tion (approximately 50%). The combination showed a istered at a constant clonidine:dexmedetomidine dose ratio of moderate (Ͻ 10-fold) but significant increase in the po- -Clon (؉ Dex)] based on the potency ratio between ago □] 1:1 nists. Note that the combination dose–response curves are plot- tency of each agonist when coadministered, the interacted as the doses of clonidine used in the presence of dexme- tion of which was statistically synergistic (isobole not ؉ detomidine. The corresponding Dex ( Clon) curve is identical shown). In ␣ AR-WT mice (fig. 5C), both clonidine and and not shown. (D) Isobolographic analysis applied to the data 2C dexmedetomidine inhibited rotarod performance at ap- from C. The y-intercept represents the ED50 for clonidine, and the x-intercept represents the ED50 for dexmedetomidine. The proximately 10-fold lower potency relative to inhibition observed combination ED50 ( ) was not significantly (P > 0.05, Œ of SP behavior. Further, the clonidine–dexmedetomi- t test) different from the theoretical additive ED50 ( ), indicat- ␣ dine combination demonstrates substantially increased ing that the interaction is additive in 2CAR-KO mice. See table 1 for ED50 values. Group sizes ranged from five to eight mice. potency (approximately 100-fold relative to each given alone) for reduction of rotarod performance. Isobolo- ␣ that in 2CAR-WT mice. These data indicate that, when graphic analysis confirmed a significant synergistic inter- given separately, neither agonist demonstrates an abso- action (isobole not shown). Figure 5D reflects a minimal ␣ Ͻ ␣ lute requirement for the 2CAR (in contrast to that seen effect ( 25%) in 2AAR-D79N mutant mice (consistent ␣ ␣ in 2AAR mutant mice), but that the 2CAR may partici- with the lack of analgesic effect); figure 5E also shows Ͻ ␣ pate in the full antinociceptive potential of the two moderate ( 50%) rotarod impairment in 2CAR-KO agonists. However, despite this moderate KO effect on mice. In summary, using the rotarod assay as a model the individual dose–response curves of clonidine and of sedation and/or motor impairment, the clonidine– dexmedetomidine, the synergistic interaction of their dexmedetomidine combination resulted in differential ␣ combination was clearly absent in the 2CAR-KO mice pharmacologic outcomes across the three lines of mice

Anesthesiology, V 110, No 3, Mar 2009 644 FAIRBANKS ET AL. Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021

Fig. 5. Clonidine (Clon)– and dexmedetomidine (Dex)–induced sedation/motor impairment. Rotarod performance was challenged ␣ by intrathecally (i.t.) administered clonidine, dexmedetomidine, or both in Institute of Cancer Research (ICR) (A), 2A-adrenergic ␣ ␣ ␣ receptor (AR) wild-type (WT) (B), 2CAR-WT (C), 2AAR-D79N (D), and 2CAR knockout (KO) mice (E). Neither clonidine ( ), dexmedetomidine (f), nor the 1:1 combination (□) exhibited greater than 40% efficacy up to the highest doses used in the substance P test. The combination did, however, result in the production of this modest efficacy at lower doses, representing significant ␣ potentiation. (B)In 2AAR-WT mice, clonidine exhibited full efficacy at 10 nmol ( ), whereas the maximum efficacy of dexmedetomidine at that dose fell short of 50% (f). The 1:1 combination (□) dose–response curve shifted significantly to the left (approximately 10-fold) relative to clonidine alone with comparable efficacy, and the interaction was found to be synergistic (isobologram ␣ not shown; P < 0.05, t test). (C)In 2CAR-WT mice, both clonidine and dexmedetomidine inhibited rotarod performance with full efficacy. The 1:1 combination (□) dose–response curve shifted significantly to the left (approximately 100-fold) relative to either drug alone with comparable efficacy, and the interaction was found to be synergistic (isobologram not shown; P < 0.05, t test). (D) ␣ f In 2AAR-D79N mice, neither clonidine ( ), dexmedetomidine ( ), nor the 1:1 combination ( ) reduced rotarod performance more ␣ f □ than 30%. (E)In 2CAR-KO mice, neither clonidine ( ), dexmedetomidine ( ), nor the 1:1 combination ( ) reduced rotarod performance more than 50%. Group sizes were five mice per group. tested in terms of relative potency and efficacy. Specifi- utility of the clonidine–dexmedetomidine combination in cally, whereas the combination significantly impaired ro- pain management or anesthesia. tarod performance in the C57Bl/6 line, it impaired motor performance only moderately in the ICR line; the effect in ␣ the 2AAR-WT line was intermediate. These results that Discussion differ across mouse lines contrast with the concurrent ␣ antinociceptive measures in that the antinociceptive po- The current study reveals that two spinally active 2- tency and efficacy and synergism were consistent (albeit of adrenergic analgesics, clonidine and dexmedetomidine, differing magnitude) across all WT mouse lines. Further interact synergistically in the production of antinocicep- analysis of other potential side effects of the combination in tion in mice. These two agonists have previously been ␣ mice and other species will be needed to determine the thought to act primarily on 2AARs to exert their various

Anesthesiology, V 110, No 3, Mar 2009 CLONIDINE–DEXMEDETOMIDINE SYNERGY IN MICE 645 physiologic effects.10,25 Because clonidine requires tion between clonidine and dexmedetomidine leaves ␣ 5 2AAR and the analgesic potency of dexmedetomidine open the possibility that the latter acts on another AR, ␣ is dramatically reduced in mice in the absence of func- such as the 2CAR. It is also conceivable that clonidine ␣ 4 ␣ tional 2AARs, the observation of synergism was an acts on 2CAR with an effect below the threshold of unexpected and novel ﬁnding. Upon further investiga- detection in our nociceptive assay. These possibilities in tion in the current study, the participation of a second turn suggested that clonidine–dexmedetomidine syn- ␣ ␣ target, likely to be 2CARs, has become apparent. The ergy may result from the participation of both 2AARs ␣ ␣ ␣ concept of 2CAR as a synergistic partner with 2AARs is and 2CARs. Two experimental tests of this hypothesis supported by previous anatomical3 and pharmacologic7 yielded concurrent results. First, cross-tolerance did not evidence. occur between clonidine and dexmedetomidine, indicat-

ing that the two agonists act at different receptors when Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 Synergistic Analgesic Pairs given as a combination. Second, clonidine–dexmedeto- ␣ Historically, synergistic analgesic partners have impli- midine synergy was not observed in 2CAR-KO mice cated the activation of two distinct receptors or receptor but did occur in WT mice. Therefore, whereas subtypes. Opioid receptor pairs with synergistic interac- clonidine and dexmedetomidine given separately by tions include MOP–DOP and MOP–KOP26; both pairings the intrathecal route seem to rely primarily on activa- ␣ involve agonists acting at separate receptor subtypes in tion of 2AAR, their spinal synergistic interaction re- ␣ the same G protein–coupled receptor family (opioid). quires the recruitment of 2CAR as well. Activity at both Others have demonstrated synergy between agonists receptors is consistent with competition binding studies, that activate receptors in different G protein–coupled where both agonists bind with comparable affinity to ␣ 2,33 receptor families: Examples include MOP and 2AR ago- both receptors. However, competition binding stud- 27,28 ␣ 4,13 nists, DOP and 2AAR agonists, and DOP and ies are incongruent with functional assays (e.g., ␣ 4 Ј ␥ 2CAR agonists. Studies evaluating interactions between guanosine 5 -O-[ -thio]triphosphate binding) in trans- agonists acting on the same opioid receptor subtype fected cell lines where dexmedetomidine has shown a 29 ␣ Ͼ ␣ Ͼ ␣ have reported only additive interactions. rank order preference for 2BAR 2CAR 2AAR and ␣ Ͼ ␣ One previous report studied the interactions between clonidine was a partial agonist at 2BAR 2AAR and ␣ 7 ␣ 34 two 2AR agonists, dexmedetomidine and ST91. The inactive at 2CAR. It is clear that in vitro binding or rationale for assessing that combination for synergy de- functional studies may not model the in vivo condition ␣ rived from observations that, whereas dexmedetomidine adequately. Furthermore, the participation of 2CAR may ␣ had been largely thought to activate 2AAR, ST91 not be at the level of direct agonist–receptor interaction ␣ seemed to be independent of 2AARs. These assertions but rather could represent an indirect contribution did not derive from binding studies because the affinities within a more complex pathway. The current study indi- ␣ of these ligands do not differ appreciably among 2AR cates that the efficacy of single agonists delivered spinally subtypes. Rather, the proposed selectivity was derived may not adequately predict the efficacy, potency, or mech- from pharmacologic studies using antagonists with dif- anism of combined agonists given spinally. ferential affinity for the three receptor subtypes.30,31 The selectivity of dexmedetomidine was subsequently vali- Interaction Studies of Sedation and dated by studies using genetically altered mice,32 but Motor Impairment ␣ ST91 did not show substantial dependence on either 2A Assessing the analgesic utility of the clonidine–dexme- ␣ 32 or 2CARs in genetically altered mice. However, the detomidine combination warrants determination of the observation that synergy was detected between dexme- effects of the combination on at least one non–analgesic- detomidine and ST91 is consistent with the participation dependent measure. Accordingly, sedation and motor of two distinct receptor subtypes. The distinct localiza- impairment were assessed using the accelerating rotarod ␣ tions of 2AAR (thought to be restricted to the spinal test immediately after antinociceptive testing. Unlike the ␣ terminals of primary afferent neurons) and 2CAR antinociceptive measure, the sedative efficacy of the (thought to be restricted to spinal neurons)3 in spinal agonists and their combination varied across the strains cord positions this pair to operate in such a synergistic studied. The individual agonists produced moderate manner. (Ͻ 50%) sedation in the outbred ICR strain, intermediate ␣ effects in the 2AAR-WT (mixed strain: C57BL/6-129sv), ␣ Clonidine–Dexmedetomidine Analgesic Synergy and pronounced sedation in the 2CAR-WT (inbred ␣ The total lack of clonidine efficacy in 2AAR functional strain: C57BL/6). Interestingly, the individual agonists knockout mice suggested that clonidine acts only at produced minimal sedation in the two mutant lines of ␣ 5 ␣ ␣ 2AARs to produce antinociception. Although the po- mice, 2AAR-D79N and 2CAR-KO, indicating that both tency of dexmedetomidine was dramatically reduced in receptor subtypes contribute to the sedative effects. The the same mice, dexmedetomidine retained analgesic ef- clonidine–dexmedetomidine combination showed a ficacy, albeit at thousandfold higher doses.4 This distinc- small sedative effect at lower doses in ICR mice, syner-

Anesthesiology, V 110, No 3, Mar 2009 646 FAIRBANKS ET AL.

␣ ␣ 54 gistic sedation in both 2AAR-WT and 2CAR-WT mice, singly delivered agents and the synergistic combina- ␣ and minimal to no sedation in both 2AAR-D79N and tions (a requisite separate study from that of the singly ␣ 55 56 2CAR-KO mice. Interestingly, a previous study of delivered agents) is imperative. The importance of ␣ 2AAR-D79N heterozygous mice revealed a clear differ- neurotoxicity evaluation of potential neuraxial therapeu- ence between the antihypertensive and sedative effects tics cannot be overemphasized.57,58 Whereas the safety of dexmedetomidine; dexmedetomidine’s cardiovascu- profile of intrathecally delivered clonidine has been pre- ␣ 53 lar effects were fully manifest in heterozygous 2AAR- viously established, the neurotoxicity of intrathecally 35 D79N mice, whereas its sedative effect was absent. delivered dexmedetomidine is largely unknown. A re- The authors attributed this difference in response to a cent evaluation59 of toxicity of epidurally delivered com- different receptor occupancy requirement for decreas- mercial dexmedetomidine formulation in rabbits found ing blood pressure versus sedation. They postulated that white matter demyelination in the spinal cord, poten- Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 ␣ partial 2AAR agonists might provide that separation of tially attributable to the pH (4.5–7.0) of the current effect in WT mice, and in fact observed a similar sepa- formulation. For the novel combination of clonidine– ration of effects in WT mice with the partial agonist dexmedetomidine to be considered useful for applica- moxonidine. Conceivably, the separation of analgesia tion, substantially more work would be needed.54 A and sedation in the outbred ICR strain results from a further consideration is that the anatomical organization similar partial agonist character of the clonidine–dexme- of ␣ ARs subtypes, although well defined in rodents, has detomidine combination. Dexmedetomidine is consid- 2 not been evaluated in human spinal cord. Differences ered a full agonist at both ␣ and ␣ ARs, whereas 2A 2C between species in receptor subtype expression pattern clonidine is considered a partial agonist at both.2 We speculate that the relation between receptor occupancy in the spinal cord could ultimately account for differ- and sedation could be a strain-dependent effect and ences in agonist combination interactions. Isobolo- account for the difference in sedative effects in the graphic analysis of a combination (fentanyl–clonidine) strains studied; however, further testing is required to well established to be synergistic in rodents did not address this hypothesis. Further study is needed to refine demonstrate statistically significant synergism in one 42 the combination to optimize clinical outcomes for either clinical evaluation ; the reason for the difference be- analgesia with moderate sedation or improved sedative/ tween rodents and humans is not clear. Regardless of anesthetic efficacy, depending on the target therapeutic these considerations, the current study reveals an unex- ␣ application. pected interaction between two 2AR agonists and sug- ␣ gests further evaluation of other 2AR agonists as poten- Clinical Relevance tially useful synergistic partners. Clinical application of interdrug synergy between G protein–coupled receptor agonists carries the potential for reduced dose and side effect profiles of drug combi- Conclusion nations compared with the drugs given alone. There is an expectation that the dose reduction enabled by a Application of interdrug synergy between G protein– synergistic interaction might reduce side effects. The coupled receptor agonists carries the potential for re- utility of clonidine as a monotherapy36–41 or combined duced dose and side effect profiles of drug combinations with spinal opioids42,43 and/or local anesthetics has compared with the drugs given alone. The potential of been studied for decades.44,45 Although the primary clin- such positive interactions encourages the continued ical use of dexmedetomidine has been as a sedative and 46,47 search for novel useful combinations. The opportunities anesthetic agent, the combination of intrathecal ␣ dexmedetomidine with bupivacaine has recently been of therapeutic application of 2AR agonists either as shown to be effective for analgesic control, comparing single agents or as combinations (particularly with opioids and local anesthetics) continues to expand with favorably with the combination of intrathecal clonidine 1 and bupivacaine.48 Further, a recent case report docu- recent clinical studies. In the current study, spinally ments the use of intrathecal dexmedetomidine com- coadministered clonidine and dexmedetomidine demon- bined with morphine to restore analgesic control in a strated a replicable and consistent synergistic interaction morphine-tolerant cancer patient.49 Therefore, both that was not predicted by previous pharmacologic stud- clonidine and dexmedetomidine produce antinocicep- ies of the agonists in genetic KO mice. The application of tion when given intrathecally both in animal mod- isobolographic analysis to this unexpected combination els4,5,50–52 and humans.36,48,49 in genetic KO mice revealed an interaction between ␣ ␣ However, before clinical application of any single 2AAR and 2CAR that would be otherwise difficult to agent or novel combination of spinal analgesics, the identify.60 Therefore, the combination of these two ago- 53 ␣ ␣ conduct of preclinical animal neurotoxicity studies nists or other coactivators of this 2A– 2CAR pair may and controlled clinical trials to establish safety of the have utility in pain management and sedative anesthesia.

Anesthesiology, V 110, No 3, Mar 2009 CLONIDINE–DEXMEDETOMIDINE SYNERGY IN MICE 647

References 30. Takano Y, Yaksh TL: Characterization of the pharmacology of intrathecally administered alpha-2 agonists and antagonists in rats. J Pharmacol Exp Ther 1992; 1. Sanders RD, Maze M: Alpha2-adrenoceptor agonists. Curr Opin Investig 261:764–72 Drugs 2007; 8:25–33 31. Takano Y, Yaksh TL: Chronic spinal infusion of dexmedetomidine, ST-91 2. Jasper JR, Lesnick JD, Chang LK, Yamanishi SS, Chang TK, Hsu SA, Daunt and clonidine: spinal alpha 2 adrenoceptor subtypes and intrinsic activity. J Phar- DA, Bonhaus DW, Eglen RM: Ligand efficacy and potency at recombinant alpha2 macol Exp Ther 1993; 264:327–35 adrenergic receptors: agonist-mediated [35S]GTPgammaS binding. Biochem 32. Stone LS, Kitto KF, Eisenach JC, Fairbanks CA, Wilcox GL: ST91 [2-(2,6- Pharmacol 1998; 55:1035–43 diethylphenylamino)-2-imidazoline hydrochloride]-mediated spinal antinocicep- 3. Stone LS, Broberger C, Vulchanova L, Wilcox GL, Hokfelt T, Riedl MS, Elde tion and synergy with opioids persists in the absence of functional alpha-2A-or R: Differential distribution of alpha2A and alpha2C adrenergic receptor immu- alpha-2C-adrenergic receptors. J Pharmacol Exp Ther 2007; 323:899–906 noreactivity in the rat spinal cord. J Neurosci 1998; 18:5928–37 33. Piletz JE, Zhu H, Chikkala DN: Comparison of ligand binding affinities at 4. Stone LS, MacMillan LB, Kitto KF, Limbird LE, Wilcox GL: The alpha2a human I-1-imidazoline binding sites and the high affinity state of alpha-2 adreno- adrenergic receptor subtype mediates spinal analgesia evoked by alpha2 agonists ceptor subtypes. J Pharmacol Exp Ther 1996; 279:694–702 and is necessary for spinal adrenergic-opioid synergy. J Neurosci 1997; 17:7157–65 34. Jansson CC, Pohjanoksa K, Lang J, Wurster S, Savola JM, Scheinin M: 5. Fairbanks CA, Wilcox GL: Moxonidine, a selective alpha2-adrenergic and Alpha2-adrenoceptor agonists stimulate high-affinity GTPase activity in a receptor

imidazoline receptor agonist, produces spinal antinociception in mice. J Phar- subtype-selective manner. Eur J Pharmacol 1999; 374:137–46 Downloaded from http://pubs.asahq.org/anesthesiology/article-pdf/110/3/638/368471/0000542-200903000-00032.pdf by guest on 26 September 2021 macol Exp Ther 1999; 290:403–12 35. Tan CM, Wilson MH, MacMillan LB, Kobilka BK, Limbird LE: Heterozygous 6. Fairbanks CA, Stone LS, Kitto KF, Nguyen HO, Posthumus IJ, Wilcox GL: alpha(2A)-adrenergic receptor mice unveil unique therapeutic benefits of partial Alpha(2C)-adrenergic receptors mediate spinal analgesia and adrenergic-opioid agonists. Proc Natl Acad SciUSA2002; 99:12471–6 ␣ synergy. J Pharmacol Exp Ther 2002; 300:282–90 36. Eisenach JC, De Kock M, Klimscha W: 2-Adrenergic agonists for regional 7. Graham BA, Hammond DL, Proudfit HK: Synergistic interactions between anesthesia: a clinical review of clonidine (1984–1995). ANESTHESIOLOGY 1996; 85 two alpha(2)-adrenoceptor agonists, dexmedetomidine and ST-91, in two sub- 655–74 strains of Sprague-Dawley rats. Pain 2000; 85:135–43 37. Eisenach JC, DuPen S, Dubios M, Miguel R, Allin D: Epidural clonidine 8. MacMillan LB, Hein L, Smith MS, Piascik MT, Limbird LE: Central hypoten- analgesia for intractable cancer pain. Pain 1995; 61:391–9 sive effects of the alpha2a-adrenergic receptor subtype. Science 1996; 273:801–3 38. Eisenach JC, Dewan DM: Intrathecal clonidine in obstetrics: Sheep studies. 9. Link RE, Stevens MS, Kulatunga M, Scheinin M, Barsh GS, Kobilka BK: ANESTHESIOLOGY 1990; 72:663–8 Targeted inactivation of the gene encoding the mouse alpha 2c-adrenoceptor 39. Mendez R, Eisenach JC, Kashtan K: Epidural clonidine analgesia after homolog. Mol Pharmacol 1995; 48:48–55 cesarean section. ANESTHESIOLOGY 1990; 73:848–52 10. Kable JW, Murrin LC, Bylund DB: In vivo gene modification elucidates 40. Eisenach JC, Lysak SZ, Viscomi CM: Epidural clonidine analgesia following subtype-specific functions of alpha(2)-adrenergic receptors. J Pharmacol Exp surgery: Phase I. ANESTHESIOLOGY 1989; 71:640–6 Ther 2000; 293:1–7 41. Eisenach JC, Rauck RL, Buzzanell C, Lysak SZ: Epidural clonidine analgesia 11. Hylden JLK, Wilcox GL: Intrathecal morphine in mice: A new technique. for intractable cancer pain: Phase I. ANESTHESIOLOGY 1989; 71:647–52 Eur J Pharmacol 1980; 67:313–6 42. Eisenach JC, D’Angelo R, Taylor C, Hood DD: An isobolographic study of 12. Hylden JLK, Wilcox GL: Intrathecal substance P elicits a caudally-directed epidural clonidine and fentanyl after cesarean section. Anesth Analg 1994; 79:285–90 biting and scratching behavior in mice. Brain Res 1981; 217:212–5 43. Siddall PJ, Molloy AR, Walker S, Mather LE, Rutkowski SB, Cousins MJ: The 13. Roerig SC, Lei S, Kitto K, Hylden JK, Wilcox GL: Spinal interactions efficacy of intrathecal morphine and clonidine in the treatment of pain after between opioid and noradrenergic agonists in mice: Multiplicativity involves spinal cord injury. Anesth Analg 2000; 91:1493–8 delta and alpha-2 receptors. J Pharmacol Exp Ther 1992; 262:365–74 44. Sites BD, Beach M, Biggs R, Rohan C, Wiley C, Rassias A, Gregory J, 14. Tallarida RJ, Murray RB: Manual of Pharmacological Calculations with Fanciullo G: Intrathecal clonidine added to a bupivacaine-morphine spinal anes- Computer Programs. New York, Springer Verlag, 1987, pp 26–231 thetic improves postoperative analgesia for total knee arthroplasty. Anesth Analg 15. Tallarida RJ: Statistical analysis of drug combinations for synergism. Pain 2003; 96:1083–8 1992; 49:93–7 45. Walker SM, Goudas LC, Cousins MJ, Carr DB: Combination spinal analgesic 16. Xie JY, Herman DS, Stiller CO, Gardell LR, Ossipov MH, Lai J, Porreca F, chemotherapy: A systematic review. Anesth Analg 2002; 95:674–715 Vanderah TW: Cholecystokinin in the rostral ventromedial medulla mediates 46. Belleville JP, Ward DS, Bloor BC, Maze M: Effects of intravenous dexme- opioid-induced hyperalgesia and antinociceptive tolerance. J Neurosci 2005; detomidine in humans, I: Sedation, ventilation, and metabolic rate. ANESTHESIOLOGY 25:409–16 1992; 77:1125–33 17. Quartilho A, Mata HP, Ibrahim MM, Vanderah TW, Ossipov MH, Lai J, 47. Kamibayashi T, Maze M: Clinical uses of ␣2-adrenergic agonists. ANESTHE- Porreca F, Malan TP Jr: Production of paradoxical sensory hypersensitivity by SIOLOGY 2000; 93:1345–9 alpha 2-adrenoreceptor agonists. ANESTHESIOLOGY 2004; 100:1538–44 48. Kanazi GE, Aouad MT, Jabbour-Khoury SI, Al Jazzar MD, Alameddine MM, 18. Moulin DE, Ling GSF, Pasternak GW: Unidirectional analgesic cross-toler- Al-Yaman R, Bulbul M, Baraka AS: Effect of low-dose dexmedetomidine or ance between morphine and levorphanol in the rat. Pain 1988; 33:233–9 clonidine on the characteristics of bupivacaine spinal block. Acta Anaesthesiol 19. Stevens CW, Yaksh TL: Studies of morphine and D-ala2-D-leu5-enkephalin Scand 2006; 50:222–7 (DADLE) cross-tolerance after continuous intrathecal infusion in the rat. 49. Ugur F, Gulcu N, Boyaci A: Intrathecal infusion therapy with dexmedeto- ANESTHESIOLOGY 1992; 76:596–603 midine-supplemented morphine in cancer pain. Acta Anaesthesiol Scand 2007; 20. Kalso EA, Sullivan AF, McQuay HJ, Dickenson AH, Roques BP: Cross- 51:388 tolerance between mu opioid and alpha-2 adrenergic receptors, but not between 50. Kalso EA, Poyhia R, Rosenberg PH: Spinal antinociception by dexmedetomi- mu and delta opioid receptors in the spinal cord of the rat. J Pharmacol Exp Ther 1993; 265:551–8 dine, a highly selective alpha 2-adrenergic agonist. Pharmacol Toxicol 1991; 68:140–3 21. Stevens CW, Monasky MS, Yaksh TL: Spinal infusion of opiate and alpha-2 51. Post C, Gordh T Jr, Minor BG, Archer T, Freedman J: Antinociceptive agonists in rats: tolerance and cross-tolerance studies. J Pharmacol Exp Ther effects and spinal cord tissue concentrations after intrathecal injection of guan- 1988; 244:63–70 facine or clonidine into rats. Anesth Analg 1987; 66:317–24 22. Paul D, Tran JG: Differential cross-tolerance between analgesia produced 52. Takano Y, Yaksh TL: Relative efficacy of spinal alpha-2 agonists, dexmedetomidine, clonidine and ST-91, determined in vivo by using N-ethoxycarbonyl- by alpha2-adrenocceptor agonists and receptor subtype selective opioid treat- ments. Eur J Pharmacol 1995; 272:111–4 2-ethoxy-1,2-dihydroquinoline, an irreversible antagonist. J Pharmacol Exp Ther 23. Aley KO, Levine JD: Multiple receptors involved in peripheral alpha 2, mu, 1991; 258:438–46 and A1 antinociception, tolerance, and withdrawal. J Neurosci 1997; 17:735–44 53. Yaksh TL, Collins JG: Studies in animals should precede human use of 24. Van der Laan JW, Van Veenendaal W, Voorthuis P, Weick G, Hillen FC: The spinally administered drugs. ANESTHESIOLOGY 1989; 70 4–6 effects of centrally acting adrenergic agonists on temperature and on explorative 54. Yaksh TL, Allen JW: The use of intrathecal midazolam in humans: A case and motor behaviour: Relation with effects on quasi-morphine withdrawal be- study of process. Anesth Analg 2004; 98:1536–45 haviour. Eur J Pharmacol 1985; 107:367–73 55. Hood DD, Eisenach JC, Tong C, Tommasi E, Yaksh TL: Cardiorespiratory 25. Guyenet PG: Is the hypotensive effect of clonidine and related drugs due and spinal cord blood flow effects of intrathecal neostigmine methylsulfate, to imidazoline binding sites? Am J Physiol 1997; 273:R1580–4 clonidine, and their combination in sheep. ANESTHESIOLOGY 1995; 82:428–35 26. Miaskowski C, Sutters KA, Taiwo YO, Levine JD: Antinociceptive and 56. Chiari A, Eisenach JC: Spinal anesthesia: Mechanisms, agents, methods, motor effects of delta/mu and kappa/mu combinations of intrathecal opioid and safety. Reg Anesth Pain Med 1998; 23:357–62 agonists. Pain 1992; 49:137–44 57. Eisenach JC, Yaksh TL: Safety in numbers: How do we study toxicity of 27. Ossipov MH, Harris S, Lloyd P, Messineo E: An isobolographic analysis of spinal analgesics? ANESTHESIOLOGY 2002; 97:1047–9 the antinociceptive effect of systemically and intrathecally administered combi- 58. Eisenach JC, James FM III, Gordh T Jr, Yaksh TL: New epidural drugs: nations of clonidine and opiates. J Pharmacol Exp Ther 1990; 255:1107–16 primum non nocere. Anesth Analg 1998; 87:1211–2 28. Ossipov MH, Harris S, Lloyd P, Messineo E, Lin BS, Bagley J: Antinocicep- 59. Konakci S, Adanir T, Yilmaz G, Rezanko T: The efficacy and neurotoxicity of tive interaction between opioids and medetomidine: systemic additivity and dexmedetomidine administered via the epidural route. Eur J Anaesthesiol 2008; 25: spinal synergy. ANESTHESIOLOGY 1990; 73:227–35 403–9 29. Pavlovic ZW, Bodnar RJ: Opioid supraspinal analgesic synergy between 60. Tallarida RJ, Porreca F, Cowan A: Statistical analysis of drug-drug and the amygdala and periaqueductal gray in rats. Brain Res 1998; 779:158–69 site-site interactions with isobolograms. Life Sci 1989; 45:947– 61

Anesthesiology, V 110, No 3, Mar 2009