Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Studies in both the mouse (Roerig and Fujimoto, 1989) and the rat (Yeung and Rudy, 1980; Siuciak and Advokat, 1989; Miyamoto et al., 1991) indicate that the concurrent administration of µ opioid receptor agonists at supraspinal and spinal sites produces antinociception in a synergistic manner. This synergistic interaction is thought to be responsible for the potent antinociception produced by systemically administered µ opioid receptor agonists, such as morphine. Although there is good agreement that supraspinal and spinal µ opioid receptor agonists interact synergistically, there is less consensus about the manner in which supraspinal and spinal  opioid receptor agonists interact. Concurrent i.c.v. and intrathecal (i.t.) administrations of  opioid receptor agonists such as [D-Ala²,D-Leu⁵]enkephalin (Roerig et al., 1991) or [D-Pen²,D-Pen⁵]enkephalin (DPDPE) (Roerig and Fujimoto, 1989) produce antinociception in an additive manner in the radiant-heat tail-flick test in the mouse. In contrast, i.c.v. and i.t. administration of DPDPE produces antinociception in a synergistic manner in a test of mechanical nociception in the rat (Miaskowski and Levine, 1992; Miaskowski et al., 1993).

Since these studies were conducted, two subtypes of the  opioid receptor, ₁ and ₂, have been described pharmacologically (Hammond, 1993; Porreca and Burks, 1993; Zaki et al., 1996). Both subtypes are implicated in the modulation of nociception in the spinal cord (Stewart and Hammond, 1993; Hammond et al., 1995) and the brain stem (Ossipov et al., 1995; Thorat and Hammond, 1997) of the rat. Although the earlier studies with DPDPE suggest that supraspinal and spinal ₁ receptor agonists interact in an additive or a synergistic manner to produce antinociception, nothing is known about the manner in which ₂ opioid receptor agonists interact. A better understanding of how  opioid receptor subtype-selective agonists interact at supraspinal and spinal sites could facilitate the development of systemically bioavailable  opioid receptor agonists as analgesics. The present study was therefore undertaken to determine how ₁ and ₂ opioid receptors in the ventromedial medulla (VMM) interact with their respective receptor subtype in the spinal cord to produce antinociception. An isobolographic analysis was conducted in which either the ₁ opioid receptor agonist DPDPE or the ₂ opioid receptor agonist DELT was administered concurrently to the VMM and the spinal cord in a fixed dose-ratio that approximated their respective ED₅₀ values at each site. Alterations in nociceptive threshold were determined by the tail-flick and hot-plate tests.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

These experiments were approved by the Institutional Animal Care and Use Committee of the University of Chicago. All procedures were conducted in accordance with the "Guide for Care and Use of Laboratory Animals" published by the National Institutes of Health and the ethical guidelines of the International Association for the Study of Pain.

Animals. Male Sprague-Dawley rats (Sasco, Kingston, NY) weighing 300 to 350 g were anesthetized with halothane and prepared with an i.t. catheter that terminated at the L4 or L5 segment of the spinal cord (Yaksh and Rudy, 1976). Rats that exhibited motor impairments such as hindlimb or forepaw paresis were euthanized. Five to 6 days later, the rats were reanesthetized with a mixture of ketamine hydrochloride (85 mg/kg i.p.) and xylazine (9 mg/kg i.p.) and implanted with an intracerebral guide cannula (26 gauge; Plastic One, Inc., Roanoke, VA) that terminated 3 mm dorsal to either the nucleus raphe magnus (NRM) or the nucleus reticularis gigantocellularis pars  (NGCp) in the VMM. The cannula was secured to the skull with stainless steel screws and dental acrylic. A 30-gauge stainless steel stylet was placed in the guide cannula to maintain its patency. Rats were housed individually after surgery under a 12-h light/dark cycle with food and water available ad libitum. Seven days elapsed before behavioral testing began. Rats received only one dose combination and were used only once in this study.

Behavioral Tests. Nociceptive threshold was assessed by the radiant heat tail-flick and 55°C hot-plate tests. In the tail-flick test (D'Amour and Smith, 1941), the rat's blackened tail was positioned under an intense light beam, and the time for the rat to remove its tail from the thermal stimulus was recorded. This test was performed twice at each time point on two different regions of the distal tail. The results of the two trials were averaged and recorded as the tail-flick latency. In the event that the rat did not withdraw its tail from the stimulus by 14 s, the test was terminated to prevent tissue damage, and the rat was assigned this cutoff latency. In the hot-plate test (Woolfe and MacDonald, 1944), the rat was placed on an enclosed copper plate heated to 55°C. The time between placement of the rat on the hot-plate and the occurrence of either a hindpaw lick or a jump off the surface was recorded as the hot-plate latency. Hot-plate latency was measured once per time period. In the absence of a hindpaw lick or a jump by 40 s, the test was terminated to prevent tissue damage, and this cutoff latency was assigned. Motor function was evaluated using the inclined-plane test (Rivlin and Tator, 1977). The tail-flick, inclined-plane, and hot-plate tests were performed in succession.

Experimental Design. Measurements of nociceptive threshold and motor competency were made before the injection of drug. Those rats that responded in 5.0 s on the tail-flick test and 15.0 s on the hot-plate test and had inclined-plane angles of 40 degrees were used in this study. Mean baseline tail-flick and hot-plate latencies among the different dose treatment groups ranged from 3.5 to 4.0 s and from 7.9 to 12.5 s, respectively. After baseline nociceptive threshold was determined, DPDPE (0.49 ng to 4.9 µg) was microinjected into the VMM followed 25 min later by an i.t. injection of DPDPE (2.3 ng to 23 µg). The intracerebral and i.t. doses of DPDPE were administered in a fixed ratio of 1:4.7 that approximated the ratio of the ED₅₀ values in mass units (i.e., nanograms or micrograms) of DPDPE at medullary and spinal sites, respectively. Tail-flick latency, hot-plate latency, and inclined-plane angle were then redetermined 45 and 60 min after the intracerebral injection. A similar paradigm was used to characterize the interaction of DELT. After determination of baseline nociceptive threshold, DELT (0.023 ng to 0.94 µg) was microinjected into the VMM, followed 10 min later by an i.t. injection of DELT (0.063 ng to 2.51 µg). The intracerebral and i.t. doses of DELT were administered in a fixed ratio of 3:8 based on the ratio of the ED₅₀ values of DELT at medullary and spinal sites, respectively. Tail-flick latency, hot-plate latency, and inclined-plane angle were then redetermined 30 and 40 min after the intracerebral injection. This order of drug administration ensured that the peak effects of DPDPE (Hammond et al., 1995; Thorat and Hammond, 1997) or DELT (Stewart and Hammond, 1993; Thorat and Hammond, 1997) at medullary and spinal sites would coincide and encompass both testing times. The agonists were administered in a fixed-dose ratio to allow characterization of the interaction between medullary and spinal sites by the isobolographic method (Roerig and Fujimoto, 1988, 1989; Tallarida et al., 1989; Tallarida, 1992b). Data on the hot-plate test were, by default, obtained at the dose-ratio determined for the tail-flick test because neither DPDPE nor DELT increased hot-plate latency in a dose-dependent manner after microinjection in the medulla. It was therefore not possible to calculate a ratio of ED₅₀ values to be administered in the hot-plate test.

Statistical Analysis. A two-way ANOVA for repeated measures was used to compare the effects of DPDPE or DELT with those of the vehicle control. The Newman-Keuls test was used for post-hoc comparisons among the individual group mean values. Dose-response relationships for DPDPE or DELT at each site alone or in combination were determined using the individual tail-flick and hot-plate latencies obtained at the time of peak effect. For concurrent injection, the dose was expressed as the total dose (VMM + i.t.) of drug administered. The ED₅₀ value was defined as the dose that produced the half-maximal possible increase in response latency. This value corresponded to 9.0 s in the tail-flick test and 25.0 s in the hot-plate test. Fieller's theorem as applied by Finney (1964) was used to determine the 95% confidence limits. The experimentally derived dose-response relationship for the total dose of drug was then compared with its theoretical dose-additive relationship by standard parallel line assay methods (Finney, 1964; Tallarida et al., 1989; Tallarida, 1992b). In addition, an isobologram was constructed using the ED₅₀ values for each agonist for the individual VMM and i.t. administrations and their 95% confidence limits (Tallarida et al., 1989). The experimentally derived ED₅₀ value of the agonist combination was then plotted on the isobologram and statistically compared with the theoretical dose-additive point (Tallarida, 1992b).

Histology. At the conclusion of testing, the rats were euthanized by CO₂ inhalation. The location and patency of the i.t. catheter were determined by direct visual inspection after a laminectomy and an i.t. injection of India ink. The brains were removed and fixed by immersion in a 4% formaldehyde and 30% sucrose solution. Transverse sections of the brain stem (25 µm) were cut on a cryostat microtome and stained with cresyl violet. The location of each microinjection site was plotted on transverse sections of the rat brain stem modified from those provided by Neurographics (Kanata, Ontario) and was verified by a person unaware of the treatment.

Drugs. DPDPE (lot no. 116H58302) and DELT (lot no. 44H08641) were purchased from Sigma Chemical Co. (St. Louis, MO) and dissolved in saline. Intracerebral microinjections were made over a 60- to 120-s period in a volume of 0.4 µl via a 33-gauge stainless steel injector that extended 3 mm beyond the tip of the guide cannula. After injection, the cannula was left in place for an additional 60 s to allow the drug to diffuse locally and to limit its diffusion up the injection track. Intrathecal injections were made over a 60-s period in a volume of 10 µl and were followed by 10 µl of saline to flush the catheter. The progress of drug delivery to supraspinal and spinal sites was monitored by the movement of an air bubble in the polyethylene tubing that connected the injector to the syringe pump.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Distribution of Microinjection Sites in VMM. Histological analysis revealed that the microinjection sites were distributed throughout the rostrocaudal extent of the NRM and NGCp. The large number of rats and different treatment groups in this study precluded presentation of all the injection sites for each treatment group; therefore, because there were no major differences in the distribution of microinjection sites among the various treatment groups, only the distribution of microinjection sites for the total dose of 0.86 ng of DELT is presented (Fig. 1). Microinjection sites located outside the NRM and NGCp included sites in the dorsal, lateral, or caudal aspects of the nucleus reticularis gigantocellularis, as well as the pyramids, inferior olive, trapezoid body, and nucleus raphe obscurus. These sites were excluded from the analysis because previous data indicated that microinjection of DPDPE or DELT at these sites did not increase tail-flick or hot-plate latency (Thorat and Hammond, 1997).

View larger version (49K): [in this window] [in a new window]   Fig. 1.   Rostrocaudal distribution of the sites in the VMM at which DELT was microinjected in a total dose combination of 0.86 ng. Numbers to the right of each section refer to the distance caudal to the interaural line. 4th vent, fourth ventricle; 7, facial motor nucleus; 7 g, genu of the seventh cranial nerve; 7t, tract of the seventh cranial nerve; dcn, dorsal cochlear nucleus; icp, inferior cerebellar peduncle; ngc, nucleus reticularis gigantocellularis; P, pyramid; V, spinal trigeminal nuclei.

Effect of Concurrent Administration of DPDPE in VMM and Spinal Cord. Total doses of 0.92 µg of DPDPE administered concurrently in the VMM and spinal cord significantly increased tail-flick latency. This effect was maximal at 45 min and persisted through 60 min after the intracerebral injection (data not shown). Total dose combinations of >27.9 µg could not be tested because it was not possible to administer >4.9 µg of DPDPE in the VMM due to its limited solubility. Figure 2 illustrates the experimentally derived dose-response relationship for the concurrent administration of the ₁ opioid receptor agonist DPDPE, as well as the theoretical additive dose-response relationship constructed for the 1:4.7 ratio of DPDPE at these sites. Although the experimentally derived dose-response relationship was situated to the right of the theoretical additive dose-response relationship (Fig. 2A), suggestive of a subadditive interaction, statistical comparison of the regression lines and their variances revealed that they did not differ significantly (P = .06), indicating that the interaction of DPDPE at medullary and spinal sites was additive. Because dose combinations of DPDPE that increased tail-flick latency beyond the 9.0-s criterion value could not be administered due to solubility limitations, an isobologram for this level of effect could not be constructed.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Dose-response relationships for DPDPE administered to the VMM (; dashed line), the i.t. space (, dashed line), and both the VMM and i.t. space in a fixed dose-ratio of 1:4.7 (; solid line). The theoretical dose-additive line for this combination is represented by the solid line (). A, tail-flick test. The slopes of the dose-response curves for DPDPE microinjected in the VMM and for i.t. administered DPDPE did not differ significantly and therefore are drawn using the common slope as described by Tallarida (1992b) When fitted to this common slope, the ED50 values and 95% confidence limits of DPDPE in the VMM and the spinal cord were 2.5 µg (1.3-4.9 µg) and 11.9 µg (4.2-34.0 µg), respectively. This common slope was then used to construct the theoretical dose-additive line for the total dose of DPDPE administered at both medullary and spinal sites; its ED50 value was 7.2 µg (2.7-11.6 µg). The experimentally derived data were also fit to the common slope. However, an ED50 value could not be calculated because the highest dose combination did not exceed the 9.0-s criterion value. B, hot-plate test. Because the microinjection of DPDPE in the VMM did not increase hot-plate latency, the theoretical dose-effect relationship for concurrent administration of DPDPE was constructed using the slope and ED50 value of i.t. administered DPDPE adjusted for the addition of an inert agent as described by Tallarida (1992b). Symbols represent the mean ± S.E.M. response latencies from 6 to 10 rats.

Effect of Concurrent Administration of DELT in VMM and Spinal Cord. Concurrent administration of DELT to the spinal cord and VMM produced an uncharacteristic dose-response relationship (Fig. 3A). Extremely low total dose combinations (86 pg to 43 ng) of DELT produced a dose-dependent increase in tail-flick latency, with a peak effect apparent by 30 min and persisting through 40 min after the intracerebral injection. Unexpectedly, the near-maximal increase in tail-flick latency produced by 43 ng was not sustained at higher doses; rather, a decrement in effect occurred. However, at total dose combinations of 0.22 to 3.46 µg, DELT again produced a dose-dependent increase in tail-flick latency. In view of the dual effect of DELT, the data for doses ranging from 86 pg to 0.043 µg of DELT and from 0.22 to 3.46 µg of DELT were separately fit by linear regression analysis and compared with the theoretical dose-additive line for the 3:8 ratio of DELT. The dose-response relationship for the very low doses of DELT was situated about 400-fold to the left of the theoretical dose-additive line and was significantly different from the dose-additive line, consistent with a synergistic interaction (P < .001). This conclusion is supported by examination of the isobologram (Fig. 4, filled circles) in which the 95% confidence limits for the ED₅₀ value of the experimental mixture do not overlap those determined for the theoretical dose-additive point. In contrast, the dose-response relationship for the total doses of DELT between 0.22 and 3.46 µg did not differ from the dose-additive line (Fig. 3A). Moreover, the experimentally derived ED₅₀ for the dose combination (Fig. 4, open circles) was superimposed on the theoretical dose-additive point in the isobologram. These data indicate an additive interaction for the higher total doses of DELT.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Dose-response relationships for DELT administered to the VMM (; dashed line) the i.t. space (, dashed line), and both the VMM and i.t. space in a fixed dose-ratio of 3:8 (; solid line). The theoretical dose-additive line for this combination is represented by the solid line (). A, tail-flick test. The slopes of the dose-response curves for DELT microinjected in the VMM and for i.t. administered DELT did not differ significantly and therefore are drawn using the common slope as described by Tallarida (1992b). When fitted to this common slope, the ED50 values and 95% confidence limits of DELT in the VMM and spinal cord were 0.63 µg (0.3-1.3 µg) and 2.0 µg (0.9-4.3 µg), respectively. The common slope was then used to construct the theoretical dose-additive line for the total dose of DELT administered at both medullary and spinal sites; the ED50 value was 1.26 µg (0.6-1.9 µg). The ED50 values for the experimentally derived data for the low and high total dose combinations of DELT administered at both medullary and spinal sites, which were also constructed using the common slope, were 0.003 µg (0.002-0.006 µg) and 1.5 µg (0.8-2.9 µg), respectively. B, hot-plate test. Microinjection of DELT in the VMM or the i.t. space did not increase hot-plate latency. These data therefore were not fit by linear regression. Symbols represent the mean ± S.E.M. of response latencies from 7 to 21 rats.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Isobologram of the ED50 values for DELT administered concurrently in the VMM and i.t. space illustrating the synergistic interaction of very low doses and the additive interaction of higher doses. Symbols on the axes represent ED50 values for DELT after administration to the VMM (abscissa) or the i.t. space (ordinate) alone. The ED50 for the experimental mixtures were obtained from the dose-response relationship for concurrent administration of total doses ranging from 86 pg to 0.043 µg of DELT () or 0.22 to 3.46 µg of DELT (); the 95% confidence limits for the experimentally derived ED50 for the low-dose combination are obscured by the symbol. The theoretical dose-additive point () for the 3:8 mixture of DELT is depicted on the line that connects the ED50 values on the abscissa and ordinate, which represents the line of additivity. Bars represent the 95% confidence limits of the ED50 values. The experimentally derived ED50 value for very low doses of DELT is significantly different from the theoretical dose-additive point (P < .001), whereas that for higher doses is not different from the theoretical dose-additive point (P > .05)

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we used an approach in which the  opioid receptor agonists were administered to medullary and spinal sites in a fixed dose-ratio. Two factors can guide the selection of the dose-ratio. The first factor is the empirical value of examining a combination that corresponds to the ratio of the equieffective (e.g., ED₅₀) doses of a drug at each site. The second factor is the variance of the theoretical dose-additive line. The selection of a dose-ratio with a small variance will facilitate detection of synergy or subadditivity. This analysis was limited to a single-dose combination that approximated the ratio of the ED₅₀ values of these drugs at each site. In the case of DELT, the variance of the dose-additive line for this ratio was near optimal. In the case of DPDPE, the variance of this ratio was not optimal. However, dose combinations with lower variance required administering a higher proportion of the total dose in the medulla; solubility limitations precluded this approach. Thus, technical limitations and the empirical value of testing equieffective doses of DPDPE at each site prevailed in the study design.

The present results indicate that concurrent administration of the ₁ opioid receptor agonist DPDPE to the VMM and spinal cord produced antinociception in an additive manner in the tail-flick test. This finding is consistent with a previous study in the mouse in which DPDPE was administered i.c.v. and i.t. in a fixed-dose ratio and antinociception was assessed by the tail-flick test (Roerig and Fujimoto, 1989). However, these results are not consistent with prior studies in the rat that concluded that i.c.v. and i.t. administered DPDPE interact synergistically to produce antinociception (Miaskowski and Levine, 1992; Miaskowski et al., 1993). One factor that may contribute to the discrepancy in findings is the use of a different measure of nociception. Miaskowski and colleagues used a modification of the Randall-Selitto paw pressure test to assess antinociception. Another factor may relate to experimental design. In these earlier studies, the researchers did not use a fixed dose-ratio analysis but rather administered increasing doses of DPDPE at one site in the presence of a fixed dose of DPDPE at the other site. This study design, in which the ratio of drug doses continuously varies, has been the subject of some debate (Tallarida et al., 1989; Tallarida, 1992a; Caudle and Williams, 1993).

The interaction of ₂ opioid receptor agonists at supraspinal and spinal sites has not been previously examined. The present findings indicate that the interaction is complex and dose dependent. Very low total dose combinations produced antinociception in the tail-flick test in a synergistic manner, whereas higher total doses produced antinociception in an additive manner. The conversion from synergy to additivity as the total dose increased was unexpected but is not without precedent. In a study of the immunosuppressive effects of µ and ₂ opioid receptor agonists using the spleen cell plaque-forming assay, the interaction of morphine and DELT converted from synergy to subadditivity as the total dose increased from femtomolar to nanomolar concentrations (Meissler et al., 1998).

At present, one can only speculate as to the mechanisms that subserve the complex interaction of medullary and spinal ₂ opioid receptors. The conversion from synergy to additivity suggests that the effects of different dose ranges of DELT are mediated by different mechanisms. One mechanism by which synergism could occur would involve the spinal release of norepinephrine. Preliminary studies from this laboratory indicate that the antinociception produced by microinjection of DELT in the VMM can be partially attenuated by i.t. administration of the ₂-adrenoceptor antagonist yohimbine (P. Banfor and D. L. Hammond, unpublished observations). Furthermore, i.t. administration of DELT with the ₂-adrenoceptor agonist dexmedetomidine produces antinociception in a synergistic manner (Grabow and Hammond, 1998). Finally, i.t. administration of yohimbine prevents the synergistic antinociception produced by coincident medullary and i.t. administration of DELT (Grabow and Hammond, 1998). Taken together, these data suggest that microinjection of low doses of DELT in the VMM evokes a release of endogenous norepinephrine in the spinal cord that interacts in a synergistic manner with exogenously administered DELT. However, as the dose of DELT increases, additional mechanisms appear to be recruited. One possible mechanism would involve an additional release of enkephalins in the spinal cord by higher doses of DELT in the VMM. Endogenously released enkephalins interact preferentially with ₂ opioid receptors in the spinal cord (Takemori and Portoghese, 1993; Tseng et al., 1995; Hammond et al., 1997) and therefore would be expected to interact in an additive manner with i.t. administered DELT. Additionally, there is new evidence that supraspinally administered morphine activates a bulbospinal pain facilitatory pathway mediated by ₁-adrenoceptors (Fang and Proudfit, 1998), in addition to a bulbospinal pain inhibitory pathway mediated by ₂-adrenoceptors (Jensen and Yaksh, 1986). The conversion to an additive interaction may reflect recruitment of an antagonistic pain facilitatory pathway (Gebhart, 1993) at higher doses of DELT.

The results with the hot-plate test provide less guidance about the interaction of supraspinal and spinal  opioid receptors, in large measure due to the limited efficacy of DELT and DPDPE at either spinal or medullary sites by themselves or after concurrent administration, as well as an inability to test high total doses due to solubility limitations in the VMM. For example, i.t. administration of DPDPE produces a dose-dependent increase in response latency (Hammond et al., 1995), but it is ineffective when microinjected in the VMM (Ossipov et al., 1995; Thorat and Hammond, 1997). Nevertheless, a theoretical dose-additive relationship for concurrently administered DPDPE could be constructed by modeling the VMM site as an ineffective agent that in effect "diluted" the i.t. administered drug (Tallarida, 1992b). Although comparison of the experimentally derived dose-response relationship to the theoretical dose-additive line did not definitively demonstrate an additive interaction, it did indicate that the interaction was not synergistic in accordance to the tail-flick test. No definitive conclusions could be reached in the case of DELT. This ₂ opioid receptor agonist did not produce a significant and sustained increase in hot-plate latency when administered at either VMM (Thorat and Hammond, 1997; but see Ossipov et al., 1995) or spinal sites (Stewart and Hammond, 1993) or when administered concurrently (this study). Although the three highest total doses produced a modest but significant increase in hot-plate latency, higher total doses could not be tested due to solubility limitations. Therefore, the existence of synergism could not be confirmed.

Support for the existence of at least two subtypes of  opioid receptor has been provided by studies that used competitive and noncompetitive antagonists of the ₁ and ₂ opioid receptor (Jiang et al., 1991; Mattia et al., 1992; Stewart and Hammond, 1993; Hammond et al., 1995) or antisense probes for the  opioid receptor (Bilsky et al., 1996) or that assessed the development of tolerance and cross-tolerance to the different prototypic agonists (Mattia et al., 1991; Sofuoglu et al., 1991). The finding that spinally and supraspinally administered DPDPE interacts in an additive manner whereas very low doses of DELT interact in a synergistic manner is additional evidence in support of the existence of different subtypes of  opioid receptor. It is unlikely that the additive interaction observed at higher total doses of DELT reflects a nonspecific action of DELT at ₁ receptors because previous studies have established that the effects of these doses of DELT in the VMM and the spinal cord are mediated by ₂ opioid receptors (Stewart and Hammond, 1993; Thorat and Hammond, 1997).

Although the existence of  opioid receptors has been recognized since the late 1970s, it is perplexing that systemically bioavailable  opioid receptor agonist analgesics have yet to be developed for clinical use. The recent introduction of SNC80 and TAN67, two structurally dissimilar nonpeptidic agonists, provided an opportunity to test the hypothesis that  opioid receptor agonists were a feasible approach to the development of potent, efficacious opioid analgesics that lacked the adverse effects associated with µ opioid receptor agonists. SNC80, which has affinity for both the ₁ and ₂ opioid receptors, was of limited potency in the tail-flick and hot-plate tests when administered systemically to the mouse (Bilsky et al., 1995). TAN67, which has preferential affinity for the ₁ opioid receptor, also exhibited limited efficacy and potency in tail-flick and hot-plate tests when administered systemically to the mouse (Kamei et al., 1995; Suzuki et al., 1995). The present finding that medullary and spinal ₁ opioid receptors interact in an additive, and not a synergistic, manner may be one basis for the relatively poor efficacy and potency of systemically administered SNC80 and TAN67. In contrast, the finding that medullary and spinal ₂ opioid receptors can interact in a synergistic manner within a limited dose range suggests that the development of systemically bioavailable ₂ opioid receptor agonists remains a feasible approach to the development of potent non-µ, non- opioid analgesics. However, in future studies of this pharmacological class of analgesic, close attention to the selection of dose will be necessary.