Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

There has been considerable effort in the scientific/medical community to develop nonopiate painkillers that maintain the efficacy of opiates against severe and chronic pain but are devoid of the opiate liabilities of respiratory depression, constipation, and dependence. The discovery that epibatidine, isolated from the skin of the frog Epipedobates tricolor (Spande et al., 1992), is an extremely potent efficacious antinociceptive agent (Spande et al., 1992; Qian et al., 1993; Badio and Daly, 1994) and the subsequent observation by our group that its activity is mediated via nicotinic receptors (Qian et al., 1993) stimulated research into the discovery of epibatidine analogs without toxic side effects. This toxicity is due to its nonspecific activity at central and peripheral nicotinic receptors (Sullivan et al., 1994; Bonhaus et al., 1995).

We synthesized several hundred analogs using the 2.2.1 azanorbornane bicycle of epibatidine as a template (Fig. 1). One of these series, containing a 1,2,4-oxadiazole ring attached to the bicycle in an exo configuration, was found to produce potent antinociception that, unlike that produced by epibatidine, is not blocked by the nicotinic antagonist mecamylamine. Two examples of this series are CMI-936 (2-exo{5-(3-methyl-1,2,4-oxadiazolyl)}-[2.2.1.]-7-azabicycloheptane) and CMI-1145 (2-exo{5-(3-amino-1,2,4-oxadiazolyl)}-[2.2.1.]-7-azabicycloheptane) (Fig. 1).

View larger version (9K): [in this window] [in a new window]   Fig. 1.   The chemical structure of ()-epibatidine, CMI-936 (2-exo{5-(3-methyl-1,2,4-oxadiazolyl)}-[2.2.1.]-7-azabicycloheptane), and CMI-1145 (2-exo{5-(3-amino-1,2,4-oxadiazolyl)}-[2.2.1.]-7-azabicycloheptane).

Having established the muscarinic nature of the antinociceptive response elicited by CMI-936 and CMI-1145, we set out to study which muscarinic subtype was involved in this antinociception. Despite five decades of research establishing a role for the muscarinic cholinergic system in antinociception (Chen, 1958; Herz, 1962; Pedigo et al., 1975; Widman et al., 1985), the subtype mediating this effect remains under debate. Dawson et al. (1991) reported that either M₁ or M₃ receptors are involved in the mouse tail-flick response to a noxious stimuli. In rats and mice, Bartolini et al. (1992) suggested that an M₁ receptor is involved in this assay. In rats injected intrathecally with muscarinic agonists, it has been suggested that muscarinic antinociception is mediated via M₁ and/or M₂ receptors (Iwamoto and Marion, 1993) or else via M₁ and/or M₃ receptors (Naguib and Yaksh, 1997). Recent evidence using the highly M₁-selective agonist xanomeline strongly argues against a role for M₁ receptors in muscarinic antinociception (Sheardown et al., 1997). Further studies using agonists with selectivity for the various muscarinic subtypes also argue against a role of M₂ and M₃ receptors in antinociception (Sauerberg et al., 1995; Shannon et al., 1997). Also arguing against a role for M₁ receptors in antinociception is the absence of M₁ receptors from the rat spinal cord (Höglund et al., 1997). In addition, in the human spinal cord M₄, but not M₁, receptors have been shown to be the predominant receptor subtype (Borenstein et al., 1996).

The identification of muscarinic receptor subtypes involved in various physiological processes has been hampered severely by the lack of antagonists selective for the five individual subtypes (Caulfield, 1993; Eglen and Watson, 1996). We thus used a combination of an antagonist (himbacine) that has selectivity for the M₂/M₄ subtypes (Waelbroeck et al., 1990; Dorje et al., 1991; Miller et al., 1992) and pertussis toxin (PTX), which is also selective for the M₂ and M₄ subtypes (Hildebrandt et al., 1983; Sternweis and Robishaw, 1984; Hulme et al., 1990), and the muscarinic M₄ toxin muscarinic toxin-3 (MT-3), which is selective for the M₄ subtype (Max et al., 1993; Jolkkonen et al., 1994). PTX selectively inhibits G_i- and G_o-coupled receptors (Hildebrandt et al., 1983; Sternweis and Robishaw, 1984). Thus, among the five muscarinic receptor subtypes, PTX inhibits M₂ and M₄ receptors that are coupled to these G proteins, but not M₁, M₃, and M₅ receptors that are coupled to different G proteins (Hulme et al., 1990). MT-3 belongs to a family of muscarinic toxins that have been isolated from the venom of Dendroaspis angusticeps (Liang et al., 1996). MT-3 has a 40-fold selectivity for the M₄ subtype over the M₁ subtype and a greater than 500-fold selectivity for the M₄ receptor over M₂, M₃, and M₅ receptors (Jolkkonen et al., 1994).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Tail-Flick Test. Female CD-1 mice weighing 20 to 30 g were obtained from Charles River Laboratories (Wilmington, MA). A commercially available tail-flick analgesia meter was used (model TF-6 Analgesia Meter; Emdie Instrument Company, Maidens, VA). The radiant heat source was set so that control mice had a tail-flick latency of 2 to 4 s. A 10-s cutoff time was used as the maximum latency to avoid damage to the mice tails. The latency of each mouse (a mean of two separate test results for each time point) was obtained at 0- (immediately before dosing), 5-, 15-, 30-, and 60-min time points after injection of compounds, and the percent maximum possible effect (% MPE) was calculated by using the formula % MPE = [(postdrug latency  predrug latency)  (cutoff time  predrug latency)] × 100.

Body Temperature. At ambient temperature, a temperature probe (Type T Thermocouple Thermometer, BAT-10; Physitemp Inc., Clifton, NJ) was inserted 1.0 cm into the rectum of mice to measure their core temperature and recorded at 0 (before drug as a control baseline), 10, 25, and 55 min after injection of compounds.

Scoring of Salivation. The salivation was noted by close visual inspection of the animal's mouth and was scored according to the following scale: 0, no sign of saliva within animal's mouth; 1, evidence of saliva in animal's mouth, but none on animal's muzzle; and 2, evidence of saliva in animal's mouth and on animal's muzzle.

Intrathecal Injection. Intrathecal injections were done freehand following the method of Hylden and Wilcox (1980). Briefly, mice were held by the pelvic girdle in one hand, as the syringe was held in the other hand at an angle of about 20° above the vertebral column. The needle was inserted into the tissue to one side of the L5 or L6 spinous process so that it slipped into the groove between the spinous and transverse process. The needle was then moved carefully forward to the intervertebral space as the angle of the syringe was decreased to about 10°. The tip of needle was inserted approximately 0.5 cm within the vertebral column. The 5 µl of solution was injected and the needle was rotated on withdrawal. Hamilton 25-µl microsyringes and 30.5-gauge needles were used in this procedure.

Drugs. CMI-936, CMI-1130, and ()-epibatidine were synthesized in the Chemistry Department of the University of Virginia. CMI-936 and CMI-1145 were prepared from the reaction of exo-2-carbomethoxy-7-azabicyclo[2.2.1]heptane with the appropriately substituted amidoxime (Carroll et al., 1993). The azanorbornane precursor was prepared as its racemic mixture either by the cyclo-addition of an N-3,5-dimethylbenzlatede pyrrole complex of pentaammineosmium (II) with methyl acrylate (Gonzalez et al., 1995) or by demethylation and ring-contraction sequence starting from tropinone. ()-Epibatidine was synthesized as previously described (Huang and Shen, 1993). Atropine, mecamylamine, and himbacine were purchased from Sigma Chemical Co. (St. Louis, MO), PTX was purchased from Research Biochemicals (Natick, MA), and MT-3 was purchased from Alexis Co. (San Diego, CA). Atropine and mecamylamine were dissolved in 0.9% saline. CMI-936, CMI-1145, ()-epibatidine, and himbacine were dissolved in dimethyl sulfoxide (DMSO). Subsequent dilutions were made in 0.9% saline such that the final concentration of DMSO in the agonist studies was less than 0.1%. DMSO (0.1%) was found to have no antinociceptive, hypothermia, or salivary effect (data not shown). PTX (50 µg/vial) was reconstituted with 1000 µl of sterile 0.01 M sodium phosphate buffer (pH 7.0) containing 0.05 M sodium chloride solution. Five microliters of this solution was injected intrathecally into each mouse such that the final dose was 0.25 µg per mouse. MT-3 (10 µg/vial) was dissolved in 50 µl of sterile 0.9% saline. Five microliters of this solution was injected intrathecally such that the final dose was 1 µg per mouse. In studies using these toxins, control animals were injected intrathecally with 5 µl of the appropriate vehicle.

Statistical Analysis. Results are given as the mean ± S.E.M. A general analysis of variance (multiway ANOVA) test using an SPSS computer program (Chicago, IL) was utilized to determine significance between control groups and antagonist/toxin-pretreated groups (multiple group comparison) for each time point in the antinociception and hypothermia data. A p value <.05 was considered significant in each case. Because the salivary data are nonparametric, a nonparametric test was used (Kruskal-Wallis H test; SPSS, Chicago, IL). A p value (corrected for ties) < .05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Antinociception of CMI-936, CMI-1145, and ()-Epibatidine. The ability of CMI-936, CMI-1145, and ()-epibatidine to elicit antinociception when injected s.c. in the mouse tail-flick assay was examined (Fig. 2). ()-Epibatidine was the most potent followed by CMI-1145 and CMI-936. All three compounds were able to elicit 100% antinociception in this assay. From these data, doses that produced between 50 and 70% of the maximum response were chosen for further study. These are 10 µg/kg for ()-epibatidine, 30 µg/kg for CMI-1145, and 50 µg/kg for CMI-936. The duration of the antinociception produced by CMI-936 and CMI-1145 was found to be greater than that for ()-epibatidine (Fig. 3). All three compounds were found to produce hypothermia, with that produced by ()-epibatidine being the most profound (Fig. 3, D-F). CMI-936 and CMI-1145, but not ()-epibatidine, were found to produce a salivary response (Table 1). The antinociception elicited by CMI-936 and CMI-1145, but not ()-epibatidine, is abolished by atropine (3 mg/kg s.c.) (Fig. 3, A-C). The salivation produced by CMI-936 and CMI-1145 (Table 1) and the hypothermia produced by CMI-936 and CMI-1145 (Fig. 3, D and E) was also abolished by atropine. By contrast, the nicotinic antagonist mecamylamine (1 mg/kg i.p.) is without effect on the antinociception elicited by CMI-936 and CMI-1145 (Fig. 3, A and B); however, it abolishes that produced by ()-epibatidine (Fig. 3C). Mecamylamine also inhibits the hypothermia produced by ()-epibatidine (Fig. 3F), but does not inhibit the hypothermia (Fig. 3, D and E) or salivation (Table 1) produced by either CMI-936 or CMI-1145. These results indicate that the antinociception, hypothermia, and salivation produced by CMI-936 and CMI-1145 are mediated via muscarinic receptors, whereas the antinociception and hypothermia elicited by ()-epibatidine are mediated via nicotinic receptors.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Antinociceptive effect of CMI-936 (), CMI-1145 (), and ()-epibatidine () in the mouse tail-flick assay. The % MPE for CMI-936 and CMI-1145 is calculated at the 30-min time point after s.c. dosing, whereas the % MPE for ()-epibatidine is calculated at the 5-min time point after s.c. dosing. Each point represents the mean ± S.E.M. of five mice.

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Effect of atropine or mecamylamine on the antinociception and hypothermia produced by 50 µg/kg s.c. CMI-936 (A and D), 30 µg/kg s.c. CMI-1145 (B and E), and 10 µg/kg s.c. ()-epibatidine (C and F) in the mouse tail-flick assay. Results are shown as responses in the absence of antagonist (), in the presence of 3 mg/kg s.c. atropine (), and in the presence of 1 mg/kg i.p. mecamylamine (). Each point represents the mean ± S.E.M. of five mice. *p < .05, **p < .01 compared with agonist responses in the absence of antagonists.

Effect of Himbacine. The M₂/M₄-selective antagonist himbacine (0.2 and 1 mg/kg s.c.) substantially inhibits the antinociceptive response elicited by both CMI-936 (50 µg/kg s.c.) and CMI-1145 (30 µg/kg s.c.) (Fig. 4,A and B). The hypothermia (Fig. 4, C and D) produced by these compounds is not inhibited by himbacine. The salivation produced by CMI-936 was unaffected by himbacine. A dose of himbacine (0.2 mg/kg), which significantly inhibited the antinociception produced by CMI-1145, was without effect on the salivation produced by CMI-1145. A dose of 1 mg/kg himbacine did produce a small inhibition of the salivary responses elicited by CMI-1145 (Table 1). The antinociception and hypothermia elicited by ()-epibatidine (10 µg/kg s.c.) is unaffected by 1 mg/kg s.c. himbacine (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Effect of himbacine on the antinociception and hypothermia produced by 50 µg/kg s.c. CMI-936 (A, C and D), 30 µg/kg s.c. CMI-1145 (B and D), in the mouse tail-flick assay. Results are shown as responses in the absence of antagonist (), in the presence of 0.2 mg/kg s.c. himbacine (), or in the presence of 1 mg/kg s.c. himbacine (). Each point represents the mean ± S.E.M. of five mice. *p < .05, **p < .01 compared with agonist responses in the absence of himbacine.

Effect of Pertussis Toxin. The ability of PTX to inhibit receptors coupled with G_i and G_o is dependent on the duration of PTX pretreatment. Several studies have shown that the effect of PTX is maximal after 7 to 12 days, at which time its effect lasts for several weeks (Hoehn et al., 1988; Galeotti et al., 1996). PTX (0.25 µg), therefore, was injected intrathecally 12 days before administration of CMI-936 and CMI-1145. The antinociception produced by CMI-936 (50 µg/kg s.c.) and CMI-1145 (30 µg/kg s.c.) is inhibited substantially by the PTX pretreatment (Fig. 5,A and B). Neither the hypothermia (Fig. 5, C and D) nor the salivation (Table 1) produced by CMI-936 and CMI-1145 is affected by the PTX treatment. The antinociception and hypothermia produced by ()-epibatidine (10 µg/kg s.c.) is also unaffected by the PTX treatment (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Effect of pertussis toxin on the antinociception and hypothermia produced by 50 µg/kg s.c. CMI-936 (A, C and D), 30 µg/kg s.c. CMI-1145 (B and D) in the mouse tail-flick assay. Results are shown as responses in the absence of the toxin () and in the presence of 0.25 µg/mouse pertussis toxin (). Each point represents the mean ± S.E.M. of 9 to 10 mice. *p < .05, **p < .01 compared with agonist responses in the absence of pertussis toxin.

Effect of MT-3 Toxin. Pretreatment with MT-3 (1 µg injected intrathecally) 20 min before the administration of CMI-936 (50 µg/kg s.c.) and CMI-1145 (30 µg/kg s.c.) significantly inhibits the antinociception produced by these agents (Fig. 6,A and B). The salivation (Table 1) and hypothermia (Fig. 6, C and D) produced by CMI-936 and CMI-1145 are unaffected by the MT-3 pretreatment. The antinociception and hypothermia produced by ()-epibatidine (10 µg/kg s.c.) is also unaffected by the MT-3 toxin (data not shown).

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Effect of MT-3 toxin on the antinociception and hypothermia produced by 50 µg/kg s.c. CMI-936 (A, C and D), 30 µg/kg s.c. CMI-1145 (B and D), in the mouse tail-flick assay. Results are shown as responses in the absence of the toxin () and in the presence of 1 µg/mouse muscarinic toxin-3 (). Each point represents the mean ± S.E.M. of 10 mice. *p < .05, **p < .01 compared with agonist responses in the absence of MT-3 toxin.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Epibatidine has been shown to be an extremely potent antinociceptive agent (Spande et al., 1992; Qian et al., 1993; Badio and Daly, 1994). The lack of separation of undesirable side effects from this antinociception (Sullivan et al., 1994; Bonhaus et al., 1995), however, preclude its development as a viable analgesic therapy. Therefore, we made several hundred analogs of epibatidine that we hoped would show separation of the antinociception from the side effects. As a result of this synthetic effort we made a series of epibatidine analogs that, in contrast to epibatidine's nicotinic mode of action, elicited potent antinociception via an activation of muscarinic receptors.

As noted in the Introduction, which muscarinic subtype mediates antinociception remains under debate (Dawson et al., 1991; Bartolini et al., 1992; Iwamoto and Marion, 1993; Sauerberg et al., 1995; Naguib and Yaksh, 1997; Shannon et al., 1997; Sheardown et al., 1997) largely because of the lack of specific agonists and antagonists for the five muscarinic subtypes. Therefore, we used a combination of an antagonist that exhibits selectivity for the M₂ and M₄ subtypes (himbacine) (Waelbroeck et al., 1990; Dorje et al., 1991; Miller et al., 1992), a toxin that exhibits similar selectivity (pertussis toxin) (Hildebrandt et al., 1983; Sternweis and Robishaw, 1984; Hulme et al., 1990), and a toxin isolated from the venom of the green mamba D. angusticeps (Liang et al., 1996) (MT-3 toxin), which has a 40-fold selectivity for the M₄ receptor over the M₁ subtype and greater than 500-fold selectivity for the M₄ receptor over M₂, M₃, and M₅ receptors (Jolkkonen et al., 1994).

Himbacine was given s.c., and as it crosses the blood brain barrier it is a useful tool to examine the role of M₂ and M₄ receptors in both peripheral and central effects of the agents used in this study. By necessity, PTX and MT-3 were given intrathecally, which allows one to determine the site of the muscarinic antinociception elicited by CMI-936 and CMI-1145.

The results with himbacine indicate that the antinociception produced by CMI-936 and CMI-1145 is a result of activation of either M₂ or M₄ receptors. By contrast, neither the salivation nor hypothermia elicited by these agents appears to be mediated via M₂ or M₄ receptors.

Further supporting a role for either M₂ or M₄ receptors in the antinociception produced by CMI-936 and CMI-1145 are the data using PTX, which allow the differentiation of the M₂ and M₄ activity of a muscarinic agonist from activity at M₁, M₃, and M₅ receptors (Hulme et al., 1990). After intrathecal injection the distribution of PTX is closely confined to the injection site (Chung et al., 1994). The data with PTX argue strongly that the muscarinic receptors mediating the antinociception are localized to the spinal cord. The lack of effect of PTX treatment on the hypothermia and salivation suggest that these effects are mediated by receptors other than spinal M₂/M₄ receptors.

To distinguish which of the M₂ or M₄ receptor subtypes is involved in the antinociception of CMI-936 and CMI-1145 it is necessary to have antagonists that show a high degree of selectivity for one subtype over the other. There are no reports to our knowledge describing compounds that exhibit this desired selectivity. Fortunately, a series of muscarinic toxins recently has been isolated from the venom of the green mamba, D. angusticeps (Liang et al., 1996). Among these is MT-3, which possesses greater than 500-fold selectivity for the M₄ receptor over the M₂ receptor (Jolkkonen et al., 1994). Because of its selectivity, this toxin has been used to show the localization of M₄ receptors to various sites, including the pain-processing region of the human spinal cord (Borenstein et al., 1996). Here, we describe the first in vivo experiments with this toxin designed to probe which muscarinic subtype is eliciting a particular response. That MT-3 pretreatment inhibited the antinociception produced by CMI-936 and CMI-1145 provides the strongest evidence for a role of the M₄ receptor in muscarinic antinociception. Because the MT-3 toxin was administered intrathecally these data again argue that these M₄ receptors are localized to the spinal cord.

Interestingly, it appears that the antinociceptive effect of CMI-1145 is more easily antagonized by himbacine and MT-3 than is the antinociceptive effect of CMI-936. One reason for this is that CMI-1145 may be mediating its antinociceptive effect solely through the M₄ receptor subtype whereas the antinociceptive effect of CMI-936 is mediated through other muscarinic subtypes in addition to the M₄ subtype.

In conclusion, we have demonstrated that certain oxadiazole analogs of epibatidine are potent antinociceptive agents in the mouse. In contrast to epibatidine, these analogs elicit their antinociception via muscarinic receptors. Furthermore, using a combination of selective toxins and antagonists we show that the M₄ receptor subtype mediates this antinociception and that these receptors are spinally located. Our data also show that by selectively targeting the M₄ receptor subtypes one should be able to elicit antinociception without eliciting the undesirable muscarinic side effects of salivation and hypothermia.