Introduction

Top Abstract Introduction Materials and methods Results Discussion References

An unexpected difference in opioid receptor selectivity occurs in the antinociceptive (tail-flick test) action of heroin between two strains of commonly used laboratory mice. Heroin given i.c.v. activates µ opioid receptors in ICR (and CD-1) mice but  opioid receptors in Swiss Webster mice (Rady et al., 1991). This difference in receptor selectivity occurs even though both sets of mice show no difference to the µ-agonist action of morphine or the  agonist action of (D-Pen^2,5)enkephalin (DPDPE). A potential pitfall associated with using randomly outbred mice such as the Swiss Webster and ICR is that the trait and genotype may vary across vendors or the strain may become commercially unavailable, as happened with Swiss Cox mice, which we had used extensively for opioid research. Also, randomly outbred mice do not provide the stable genotype (within or across suppliers) that is necessary for a behavioral genetic analysis, nor do they provide a consistent neurobiological pool that can be assessed across laboratories and time. The advantage of inbred strains is that much is known about genetic composition and origin of the strains (origins of inbred). In addition, in many cases much is known about the behavioral and neurochemical profile of the genotype (Belknap and O'Toole, 1991). Furthermore, the ability to determine genetic loci (and possibly the transcribed proteins) that account for genotype-dependent variations in response to a drug makes genetic control both a short- and long-term advantage (Crabbe et al., 1994). Because previous research had suggested that genotype significantly influences the pharmacological effects of heroin, one purpose of the current study was to determine the degree to which genotype governs opioid-receptor selectivity for heroin in inbred mice.

A related purpose of this series of studies was to identify genotypes in which heroin acts through different opioid receptors. Identification of genotypes with distinct pharmacological reactions to heroin has several advantages. First, it clearly establishes the degree to which genotype determines the pharmacological actions of heroin. Second, it characterizes the pharmacology of heroin in a broad population. Third, it provides a system in which the biochemical and molecular mechanisms underlying agonist receptor selectivity can be further analyzed. Finally, it establishes a database to explore the pharmacological basis of vulnerability to drug use and abuse.

The most direct way to evaluate the type of opioid receptor involved is to use selective opioid receptor antagonists and perform major ED₅₀ studies for heroin. Initial single-dose studies with opioid antagonists would provide information for the design of the subsequent ED₅₀ experiments. However, such initial results would not provide information beyond those derived from the major studies. On the other hand, intial studies on descending pathways of heroin-induced antinociception would provide complementary results useful beyond those from the ED₅₀ studies. The tail-flick test depends on a spinal reflex [it remains intact after transection of the cervical spinal cord (Wang et al., 1994)], which can be inhibited by activating descending antinociceptive pathways to the spinal cord. The loci of action of opioids in the brain, the opioid receptors there, and the descending systems with neurotransmitters in the spinal cord have been reviewed by Yaksh and Malmberg (1994). Differential activation of spinopetal antinociceptive pathways occurs after i.c.v. administration of opioid receptor agonists (Fig. 1). In mice, activation of  receptors in the brain leads to antinociception modulated by spinal serotonin receptors (Ho and Takemori, 1989). Spinal serotonin and the ₂-action of norepinephrine are involved in the analgesic action of µ agonists given i.c.v. (Hylden and Wilcox, 1980; Arts et al., 1991). Activation of  receptors in the brain involves GABA receptors in the spinal cord (Holmes and Fujimoto, 1994; Rady and Fujimoto, 1995, 1996). Supraspinal opioid-induced antinociception can be selectively inhibited by IT administration of antagonists to the neurotransmitter mediating each descending system: methysergide for serotonin, yohimbine for alpha₂ noradrenergic and bicuculline for GABA_A receptors. In Swiss Webster mice, where i.c.v. heroin acts on  receptors, IT bicuculline shifts the dose-response curve for heroin to the right in parallel fashion (Rady and Fujimoto, 1995), while IT yohimbine and methysergide produce parallel rightward shifts for i.c.v. morphine (Suh et al., 1989). The initial results were used in designing more extensive experiments, wherein heroin dose-response relationships were generated after i.c.v. administration of opioid receptor antagonists: -funaltrexamine (-FNA) for µ, naltrindole for , and N-BNI for  receptors.

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Scheme for the pharmacological determination of the receptor selectivity of heroin given i.c.v. based on inhibition of descending (serotonergic, GABAergic, noradrenergic) systems. The various antagonists with dose and route of administration are given in the ellipse, and the lines connecting downward indicate the progression of the analyses. The inbred strains that were ultimately classified as groups A, B, and C are indicated in the rectangles.

	Materials and Methods

Top Abstract Introduction Materials and methods Results Discussion References

Animals. Adult male AKR/J, BALB/cByJ, C57BL/6J, C3H/HeJ, CBA/J, and DBA/2J mice (Jackson Laboratories, Bar Harbor, ME), 60 to 120 days old and weighing approximately 23 to 30 g at the start of the experiment, were used. Animals were housed in groups of three to five in a temperature-controlled room (21°C) with a 12-h light/dark cycle (lights on at 7:00 AM). Animals had free access to Purina Laboratory Chow and tap water at all times. The studies were conducted in accordance with the Guide for Care and Use of Laboratory Animals provided by the National Institutes of Health and adopted by National Institute on Drug Abuse and were in compliance with the Institutional Animal Care and Use Committee (Animal Studies Subcommittee).

Antinociceptive Response. The radiant heat tail-flick test (D'Amour and Smith, 1941) was used to measure antinociception. The lamp intensity was set to provide a predrug response time of 2 to 4 s, and a cut-off time of 10 s was used as the maximal response. The postdrug tail-flick latency was converted to percent maximum possible effect (% MPE) as calculated by Dewey et al. (1970): % MPE = (postdrug time  predrug time) × 100/(10  predrug time).

Intracerbroventricular and IT Drug Administration. The basic protocol involved administration of heroin or other opioid agonists i.c.v. to inhibit the tail-flick response. The approach to peak action of i.c.v. agents was assumed to be 10 min after administration based on previous studies in outbred mice (Rady et al., 1991, 1994a, b; Holmes and Fujimoto, 1994; Rady and Fujimoto, 1995, 1996). This 10-min time favored heroin as the primary antinociceptive agent rather than its metabolites, 6-monoacetylmorphine and morphine (Rady et al., 1991; Rady and Fujimoto, 1995, 1996). Intracerbroventricular administration was by the method of Haley and McCormick (1957) under light halothane anesthesia. In later experiments opioid antagonists were administered i.c.v. along with heroin. IT administration of antagonists (bicuculline, methysergide, yohimbine) was by the method of Hylden and Wilcox (1980).

Single-Dose Studies Involving Descending Systems. The most direct way to determine the opioid receptors involved in heroin analgesia would be to use selective opioid receptor antagonists i.c.v. and determine the shifts they produce in the dose-response curves for heroin. However, -FNA (a nonequilibrium µ antagonist), N-BNI (a  receptor antagonist), and naltrindole (a  receptor antagonist) have long duration of action (Ward et al., 1982; Takemori et al., 1988; Horan et al., 1992) making reuse of the mice difficult. Reuse was desirable because some inbred strains were not readily available. The use of the nonopioid antagonists for the neurotransmitters of the descending systems (see Introduction), which had short duration of action (Suh et al., 1989; Arts et al., 1991; Holmes and Fujimoto, 1994) allowed the mice to be used more than once. The mice were used a total of 3 times with 5 to 7 days between uses. The treatments were randomized and no order effects were found. Previously Mickley et al. (1990) reported results from C57BL/6J mice reused 3 days after i.c.v. administration of µ and  receptor agonists.

Dose-Response Relationship Experiments for i.c.v. Heroin with Opioid Receptor Antagonists (ED₅₀ Determinations). Additional pharmacological determination of receptor selectivity for i.c.v. heroin was based on the use of opioid receptor antagonists given i.c.v. These mice were experimentally naive and were not reused. The results from the single-dose experiments with the descending systems indicated that all three of the known opioid receptor types (µ, , and ) had to be considered. The opioid antagonists used were -FNA for µ, naltrindole for , and N-BNI for  receptors. Doses were based on our previous publications (Rady et al., 1991, 1994a, b). Naltrindole and N-BNI were administered i.c.v. at 10 min in the same solution with heroin; -FNA was given i.c.v. as a 24-h pretreatment.

Statistical Analysis. Student's t test (comparison between two groups) and analysis of variance (ANOVA) followed by Dunnett's test (comparison of each group to a control group) were used for determining significant differences between group means as indicated by a P  .05 (Steel and Torrie, 1960). The ED₅₀ values with the 95% confidence intervals were determined and compared for significant differences (P  .05) according to the method of Litchfield and Wilcoxon (1949) as described for the computerized version by Dewey et al. (1970) and used previously (Roerig and Fujimoto, 1989).

Source of Drugs and Drug Solutions. DAMGO (Peninsula, Belmont, CA); heroin hydrochloride (National Institute on Drug Abuse, Rockville, MD); morphine sulfate (Mallinckrodt Chemical Works, St. Louis, MO); (+)-bicuculline, DPDPE and yohimbine hydrochloride (Sigma Chemical Co., St. Louis, MO); U50,488H (UpJohn, Kalamazoo, MI); methysergide maleate (Sandoz, Berne, Switzerland), -FNA and N-BNI (RBI, Natick, MA); naltriben and BNTX were obtained from Takemori and Portoghese (University of Minnesota, Minneapolis, MN). The doses of the drugs were for the forms stated above. Drugs were dissolved in a 0.9% sodium chloride solution or in a 0.01% Triton X-100 in 0.9% sodium chloride solution (DPDPE, DAMGO). Slight heating was used to dissolve bicuculline. Doses and times of administration of the drugs were taken from our previous publications and are given with each experiment under Results.

	Results

Top Abstract Introduction Materials and methods Results Discussion References

The results are grouped according to the receptor selectivity found for heroin. In group A, heroin acted on  receptors, in B on µ receptors, and in C on  receptors.

Single-Dose Studies Involving Descending Systems. Group A ( Receptor Response). The results in Fig. 2 for the C57BL/6J mice indicate that a 3-µg dose (used as the standard dose in all strains except AKR/J) of heroin given i.c.v. 10 min before the tail-flick test produced about an 80% MPE. This antinociception was inhibited by IT bicuculline (0.5 µg), but not by methysergide (1 µg), given 5 min before the tail-flick test. These results indicated that the response involved supraspinal  and not µ or  receptors. In the test for the subtype of  receptor (see Fig. 3), the ₁ receptor antagonist BNTX, 1 pmol (0.465 ng) given with the heroin, inhibited the response while naltriben, 25 pmol (12.8 ng), a ₂ receptor antagonist, did not. These doses of BNTX and naltriben were shown previously in Swiss Webster mice to be effective in inhibiting the ₁ and ₂ response, respectively (Rady et al., 1994). Thus, the i.c.v. heroin receptor response involved ₁ receptors. The methysergide treatment was effective in inhibiting the antinociceptive action of i.c.v. morphine (Table 1). Therefore, the serotonergic pathway was present but was not involved in heroin action.

View larger version (62K): [in this window] [in a new window]   Fig. 2.   The ability of IT bicuculline and methysergide (step 1) to inhibit i.c.v. heroin in six strains of inbred mice determined with the tail-flick test. Heroin, 3 µg (6 µg in AKR/J) was given i.c.v. to all groups 10 min before the tail-flick test. The other treatments given IT at 5 min before the tail-flick test are indicated vertically for each group. Heroin antinociception was attenuated by IT bicuculline only in C57BL/6J mice to implicate  receptors. IT methysergide inhibited heroin antinociception in CBA/J, DBA/2J BALB/cByJ, and AKR/J mice, suggesting either µ or  receptor mediation. Additional analyses were performed in the second step, next figure. The bars represent the mean % MPE, with the S.E.M. indicated by the vertical line at the top of the bar with the number of mice in each group given along side. * Indicates that the mean was significantly different from the control group mean using Dunnett's test (more than one group compared with control); P  .05.

Group B (µ Receptor Response). CBA/J, DBA/2J, and C3H/HeJ. Bicuculline did not affect heroin action in CBA/J, DBA/2J, and C3H/HeJ, indicating that  receptors were not involved (Fig. 2). Inhibition by methysergide indicated that in CBA/J and DBA/2J either µ or  receptors were involved. The effectiveness of IT yohimbine (Fig. 3) indicated µ receptor participation in the CBA/J response. The lack of response of DBA/2J to N-BNI eliminated the alternative of the  receptor and classified DBA/2J as µ responders. The lack of response to IT yohimbine (Fig. 3) was considered to be a reflection of the different amounts of participation between serotonergic and noradrenergic systems in the µ response. In this regard, the response in C3H/HeJ was insensitive to methysergide (Fig. 2) but sensitive to yohimbine inhibition (Fig. 3), making this strain µ responders. The positive yohimbine response eliminated  receptors. Thus, µ receptors were involved in the action of heroin in CBA/J, DBA/2J, and C3H/HeJ mice.

View larger version (64K): [in this window] [in a new window]   Fig. 3.   Second step in sequential test to differentiate between opioid receptors suggested by the previous test. The antinociceptive response to i.c.v. heroin in C57BL/6J was reduced by coadministration of 1 pmol of BNTX, but not 25 pmol of naltriben, implicating involvement of a 1 opioid receptor response. Inhibition of antinociception in CBA/J and C3H/HeJ by IT yohimbine implicated µ receptor mediation. The lack of inhibitory action by i.c.v. N-BNI in DBA/2J indicated that K receptors were not involved so that the inhibitory action of IT methysergide in the previous experiment (see Fig. 2) was due to µ rather than  receptors. In BALB/cByJ and AKR/J, the antagonistic effect of N-BNI and lack of effect of IT yohimbine and -FNA given 24 h before the tail-flick test implicated  receptor mediation. * Indicates that the mean was significantly different (P  .05) from the control group mean using Dunnett's test of Student's t-test (one group compared with control).

Group C ( Receptor Response): BALB/cByJ, AKR/J. The action of i.c.v. heroin was not inhibited by IT bicuculline but was inhibited by IT methysergide in both BALB/cByJ and AKR/J (Fig. 2) The antinociceptive action of heroin was also not inhibited by administration of yohimbine IT (Fig. 3). Because, µ receptor action is not always associated with the descending noradrenergic activation, lack of yohimbine effect did not necessarily rule out µ receptor action in these strains. A 24-h pretreatment with 25 ng of i.c.v. -FNA in BALB/cByJ was not effective (Fig. 3); thus, µ receptors were not involved. This negative finding alerted us to the need to consider  receptors, even though there was no hint previously that heroin could have  agonist activity. Strikingly, i.c.v. administration of N-BNI was effective in both BALB/cByJ and AKR/J strains (Fig. 3). Therefore, heroin action was due to activation of  receptors, which are associated with the descending serotonergic system. Note that in AKR/J, the dose of heroin was 6 µg i.c.v., compared with 3 µg in all the previous experiments. These mice were less sensitive to i.c.v. heroin. IT bicuculline inhibited the antinociceptive action of i.c.v. DPDPE (Table 1) in BALB/cByJ and AKR/J, indicating that  receptors and the descending GABA systems were present but not acted on by heroin.

Intracerbroventricular Heroin ED₅₀ Determinations and Effect of Opioid Antagonists. The ED₅₀ values and 95% confidence intervals derived for i.c.v. heroin alone or in the presence of the opioid antagonists for the six strains are given in Table 2. The dose-response curves are given in Fig. 4. These experiments give more directly focused results than those derived from the initial single-dose experiments.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Dose-response relationship to i.c.v. heroin and the effect of opioid receptor antagonists. Heroin was given at different doses, i.c.v., 10 min before the tail-flick test (open circle). Intracerbroventricular -FNA, 25 ng, treatment (solid circle) was given 24 h before the heroin. The i.c.v. naltrindole (10 µg, black triangle) and N-BNI (10 µg, black square) were given together with the heroin. For each point on the graph, 7 to 8 mice were typically used; while for one quarter of the points, 6 mice were used. The thin black line always designates heroin alone (control), heavy black line is for the antagonist that produced the greatest shift to the right, and the dotted lines are for the other two antagonists, which produced no significant shift. The ED50 values are given in Table 2.

Group A. Treatment with naltrindole shifted the dose-response curve of heroin in C57BL/6J mice to the right (Table 2 and Fig. 4). Twenty-four hour pretreatment with i.c.v. -FNA or administration of N-BNI along with heroin did not affect the ED₅₀ value. Thus, the response to heroin in C57BL/6J mice involved  but not µ or  opioid receptors.

Group B. The ED₅₀ values and dose-response curves for heroin in DBA/2J and CBA/J mice indicated that 24-h -FNA treatment inhibited the analgesic response of i.c.v. heroin (Table 2 and Fig. 4). N-BNI and naltrindole treatments had no effect. These two strains gave µ receptor responses to heroin. The results for the C3H/HeJ mice did not give a clear differentiation. Even though the data analysis in Table 2 indicated that -FNA had a small but significant inhibitory effect on heroin antinociception, it was not significantly different from the heroin response in the presence of N-BNI. Thus, these results gave weak support to the conclusion that C3H/HeJ mice were µ responders to heroin.

Group C. Both the BALB/cByJ and AKR/J strains showed a robust inhibition of the heroin response by N-BNI (Table 2 and Fig. 4). Naltrindole and -FNA treatments had no effect on the heroin ED₅₀ values. Thus, both strains were designated as  receptor responders to heroin.

	Discussion

Top Abstract Introduction Materials and methods Results Discussion References

Previous findings in outbred mice suggest that genotype significantly influences the pharmacology of heroin-induced antinociception (Rady et al., 1991, 1994b). The current studies confirmed these findings and characterized the degree to which genotype governs opioid-receptor selectivity in standardized inbred mice. Because the inbred strains used in these experiments were studied under identical conditions (environments), differences observed in receptor selectivity and descending control mechanisms were due to genetic differences across individual strains. The initial single-dose experiments to determine descending pathways require cautious interpretation, particularly when an antagonist did not affect the heroin response. Larger doses of heroin might activate more than one system. When the MPE is near 100%, an antagonist selective for the low-dose effect of heroin might be ineffective because of the presence of the second action of heroin. Such concern is diminished in part if the results from the descending systems and the ED₅₀ determinations lead to the same conclusion. The results of both studies are discussed together.

Agonist Action: C57BL/6J. In these mice, IT bicuculline but not methysergide inhibited the response to heroin. This was consistent with i.c.v. heroin-activating supraspinal  receptors to activate spinal GABA_A receptors. The robust shift in the heroin ED₅₀ value produced by i.c.v. naltrindole and lack of effect of N-BNI or -FNA treatment provided direct evidence for a predominant  receptor response. Additional differentiation with i.c.v. BNTX suggested that i.c.v. heroin acted on supraspinal ₁ receptors. These findings are similar to those in Swiss Webster mice (Rady and Fujimoto, 1995; Rady et al., 1991).

µ Agonist Action: CBA/J, DBA/2J, C3H/HeJ. The antinociceptive responses to i.c.v. heroin in these three inbreds were ascribed predominantly to supraspinal µ receptor involvement. The descending system could be serotonergic (DBA/2J), noradrenergic (C3H/HeJ), or both (CBA/J). Different µ agonists show heterogeneity for activation of serotonergic and noradrenergic systems (Arts et al., 1991). Even though IT yohimbine inhibited i.c.v. heroin-induced analgesia in C3H/HeJ mice, the µ receptor assignment was weakened because the shift of the heroin ED₅₀ value produced by -FNA was very small. Increasing the dose of -FNA did not improve this result (data not given). The ED₅₀ results for the other two strains agreed with the findings of the single-dose studies.

Agonist Action: AKR/J, BALB/cByJ. Heroin antinociception was inhibited by IT methysergide in both strains. Intracerbroventricular N-BNI shifted the dose-response curve of heroin well to the right, while i.c.v. -FNA and naltrindole had no effect. Because the tail-flick test is generally not sensitive to  agonists (Taber, 1974), obtaining analgesia for i.c.v. heroin as  receptor responses was unusual. An ED₅₀ value of 2 µg (about 5 nmol) suggests that heroin is about 4 times more potent than U50,488H as a  agonist. Because  agonists have dysphoric effects, it would be of interest to determine if the  agonist analgesic action of heroin were not limited by dysphoric actions.