Clara C Faura, Instituto de Neurociencias, Universidad Miguel Hernández – CSIC, Avda. Ramón y Cajal s.n., San Juan de Alicante, Alicante 03550, Spain. E-mail: firstname.lastname@example.org
BACKGROUND AND PURPOSE The perception of pain and its inhibition varies considerably between individuals, and this variability is still unexplained. The aim of the present study is to determine whether functional interactions between opioid receptors are involved in the inter-individual variability in the sensitivity to μ-opioid receptor agonists.
EXPERIMENTAL APPROACH Anti-nociceptive tests, radioligand binding, stimulation of [35S]GTP–γ–S binding, inhibition of cAMP production and co-immunoprecipitation experiments were performed in two strains of rat (Sprague–Dawley bred at our university – SDU – and Wistar) that differ in their sensitivity to opioids.
KEY RESULTS The increased anti-nociceptive potency of µ-opioid receptor agonists in SDU rats was reversed by the δ-opioid receptor antagonist, naltrindole. Inhibition of the binding of [3H] naltrindole by µ-opioid receptor agonists was different in brain membranes from SDU and Wistar rats. Differences were also evident in the effect of δ-opioid receptor ligands on the binding of [35S]GTP-γ–S stimulated by µ-opioid receptors agonists. No strain-related differences were detected in spinal cord membranes. The potency of morphine to inhibit cAMP production in brain membranes varied between the strains, in the presence of deltorphin II and naltrindole. Co-immunoprecipitation experiments demonstrated that δ-opioid receptors were associated with μ-opioid receptors to a higher extent in brain synaptosomal fractions from SDU than in those from Wistar rats.
CONCLUSIONS AND IMPLICATIONS There was increased supraspinal cross-talk between μ and δ-opioid receptors in SDU, as compared with Wistar rats. This was related to an enhanced sensitivity to anti-nociception induced by µ-opioid receptor agonists.
Opioids are the most powerful analgesics available currently for patients, yet their use and efficacy remains limited by undesirable side effects and the inter-individual variability found in opioid responses. The perception of pain and its inhibition varies considerably between individuals, both in humans (Dionne et al., 2005) and in other animals (Mogil et al., 1996; Más et al., 2000). This variability is still unexplained. From the clinical point of view, it is important to recognize that optimal pain control may be achievable through understanding the molecular-genetic mechanisms, yielding individualized analgesic medications and dose regimens based upon each person's genetic endowment (Kim and Dionne, 2009). Several genetic models have been identified or produced using rodents, in an attempt to investigate the mechanisms that underlie variation between individuals in their response to opioid drugs. Differences have been described in nociceptive sensitivity (Hoffman et al., 1998) and in the anti-nociceptive effect of opioids (Hoffman et al., 1998; Morgan et al., 1999; Más et al., 2000), as well as in other actions of opioids (Hoffman et al., 1998; Más et al., 2000).
The effects of opioid drugs and endogenous opioid peptides are mediated by μ-, δ-, and κ-opioid receptors found in the CNS and other tissues (receptor nomenclature follows Alexander et al., 2011). Opioid receptors belong to the family of GPCRs and their multiplicity provides a basis for explaining the complex pharmacology of opioids.
At present, the existence of interactions between opioid receptors is widely assumed. There is indirect evidence that opioid receptors do not necessarily act independently from each other. The existence of opioid receptor complexes was reported more than 30 years ago from radioligand binding and anti-nociception experiments (Vaught and Takemori, 1979). The cross-talk between μ-opioid receptors and δ-opioid receptors is documented, mainly from the observation that δ receptor agonists modulate μ receptor–mediated analgesia (Vaught et al., 1982). Allosteric coupling between μ and δ-opioid receptors has been also postulated (Rothman et al., 1993). Furthermore, immunohistochemical studies have demonstrated co-localization of opioid receptors in some neurons (Ji et al., 1995), and cross-linking and co-immunoprecipitation studies have suggested the existence of a receptor complex both at spinal and supraspinal levels (Schoffelmeer et al., 1990; Gomes et al., 2004; Garzón et al., 2005), opening the possibility for physical interactions between the receptors. It has been also shown that μ- and δ-opioid receptors coexist in a complex with a series of signalling elements such as G-proteins and RGS proteins (Garzón et al., 2005). Functional interactions between μ- and δ-opioid receptors have been demonstrated in knockout models. Mice lacking μ-opioid receptors show reduced anti-nociception and the absence of respiratory depression induced by δ-opioid receptor agonists (Matthes et al., 1998). In mice lacking δ-opioid receptors, δ receptor agonists retain supraspinal analgesia, suggesting the existence of a second δ-receptor like analgesic system or μ receptor-dependent anti-nociception (Zhu et al., 1999; Scherrer et al., 2004). Also, δ-opioid receptor knockout mice do not develop analgesic tolerance to morphine (Zhu et al., 1999). Co-expression of δ- and μ-opioid receptors enhances the signalling of both receptors, suggesting that there is functional interaction between them (Snook et al., 2008). The putative coordinated action of the opioid receptors, which may have therapeutic implications, needs to be clarified.
The aim of the present work was to assess whether the extent of functional interaction between μ- and δ-opioid receptors could determine the sensitivity to anti-nociception induced by μ-opioid receptor agonists. We addressed this question using a number of in vivo and in vitro approaches in two strains of rats – Sprague–Dawley bred at our university (SDU) and Wistar – that differ in their sensitivity to morphine. Our findings demonstrate that the sensitivity to the anti-nociceptive effect of μ-opioid receptor agonists was related to the extent of the interaction between μ- and δ-opioid receptors at a supraspinal level.
All animal care and experimental procedures were approved by the Institutional Animal Care and Use Committee and followed guidelines regarding ethical standards for the experimental investigation of pain in animals (Zimmermann, 1983). All experiments were carried out on adult male rats that were 12–15 weeks old. The strains used were SDU rats, derived from a line bred at our University, and Wistar rats that were purchased commercially (Harlam, Barcelona, Spain). The animals were housed in clear plastic cages, three to four rats per cage, and maintained on a 12 h light/dark cycle, with sawdust bedding. Food and water were provided ad libitum, and the humidity and temperature were stabilized (20–22°C). The experiments were carried out between 1000 and 1500 h.
The tail flick test was performed as described by Más et al. (2000). The light intensity (55–60°) was adjusted such that the baseline latencies ranged between 3 and 5 s, and the individual cut-off was calculated using the formula: cut-off = baseline latency × 10/4.5 (Menendez et al., 1993). The hot plate response was assessed as previously described (Le Bars et al., 2001). The heated surface of the plate was kept at a temperature of 55 ± 0.5°C. The reaction times of the first evoked behaviour regardless of whether it was paw licking or jumping were measured, and the cut-off was 30 s. Dose–response curves for morphine, fentanyl, methadone and 3-d were obtained for each animal by using a cumulative-dosing procedure (Paronis and Holtzman, 1991). After habituation and baseline trials, each animal was injected s.c. or i.p. with a low dose of morphine, fentanyl, methadone or 3-d, such that the cumulative dose was increased by 0.15 log units. Anti-nociception was evaluated by hot plate and tail flick tests 30 min after each cumulative s.c. dose of morphine (3.3–27 mg·kg–1 for the hot plate response and 0.3–9.6 mg·kg–1 for the tail flick test) and by tail flick test alone 30 min after each cumulative dose of methadone (0.85–4.8 mg·kg–1), 15 min after each dose of fentanyl (2.5–57 μg·kg–1), and 5 min after each dose of 3-d (0.15–1.7 mg·kg–1). This procedure was continued until the rat showed no nociceptive response within the cut-off period. Non-cumulative dose–response curves were obtained for SNC80 (6.8–13.5 mg·kg–1). The effect of the selective δOR antagonist naltrindole on the morphine and fentanyl dose–response curves was determined by administration of naltrindole by s.c. injection 20 min before different doses of morphine or fentanyl were administered.
Results are expressed as means ± SEM. The anti-nociceptive responses were expressed as a percentage of the maximum possible effect (%MPE), which was calculated as: %MPE = 100 ×[(test latency – baseline latency) / (cut-off time – baseline latency)]. The anti-nociceptive dose–response curves were fitted to the logit–log equation, and the ED50 was estimated with Prism (GraphPad Software Inc., La Jolla, CA). The dose–response curves of nociceptive tests were compared using a two-way anova, with strain or co-administration of naltrindole as the between-groups factor and dose as the repeated-measures factor. This was followed by the Bonferroni test for multiple comparisons. The level of significance was set at P < 0.05.
For each batch, pooled membranes from the whole brain (minus cerebellum) or spinal cord from six rats were prepared (Fang et al., 1994). Protein concentrations were determined using the Bradford method (Bradford, 1976). The binding of [3H]naltrindole (specific activity 60 Ci·mmol–1) to the microsomal brain membranes was performed as described by Fang et al. (1994). In the inhibition experiments, 0.1 nM [3H]naltrindole was used. The binding of [3H]DAMGO (specific activity 60 Ci·mmol–1) was performed essentially as described by Raynor et al. (1995). In the inhibition experiments, 1 nM [3H]DAMGO was used. Non-specific binding of both radioligands was determined in the presence of 10 μM naloxone. Experimental data were expressed as specific binding (fmol·mg–1 protein). The data on saturation binding were fitted to a one binding-site saturation model with Prism. The results of the inhibition experiments were fitted to one- and two-binding sites models using the programme Prism, and F-tests were performed to determine whether the experimental data fitted significantly better to a model with one or two binding sites.
The binding of [35S]GTP-γ–S (specific activity 1000 Ci·mmol–1) to pooled membranes from brain or spinal cord from both strains, prepared as above, was investigated as previously described (Sim et al., 1995), with minor modifications. The basal level of binding was determined by incubation of samples in the absence of agonists. Non-specific binding was determined in the presence of 100 μM unlabelled GTP-γ–S. To study interactions between the μ- and δ-opioid receptor ligands, increasing concentrations of δ receptor ligands were used in the presence or absence of a fixed dose of the μ receptor agonists, DAMGO (30 nM for deltorphin II and TIPPΨ and 3 μM for naltrindole) or morphine (30 nM for deltorphin II and TIPPΨ and 0.5 μM for naltrindole).
Experimental data on [35S]GTP-γ–S were expressed as the percentage of the basal binding. Results were fitted to appropriate binding models with the programme Prism. When necessary, F-tests were performed to determine whether the experimental data fitted significantly better to one- or two-binding sites model.
For the cAMP assay, membranes were prepared from the whole brain (minus cerebellum), and cAMP was extracted from the tissue homogenate essentially as described by Izenwasser et al. (1993). The ability of morphine or deltorphin II to inhibit the adenylyl cyclase activity stimulated by 50 μM forskolin was evaluated by incubation of aliquots of the brain homogenate (protein concentration of 40 μg/60 μL final volume) with increasing concentrations of morphine or deltorphin II (0.01–100 μM, in both cases), for 5 min at 30°C. The basal activity of adenylyl cyclase was determined by incubation of the membranes in the absence of forskolin. To study the interactions of δ-opioid receptor ligands with morphine, we used the same protocol to evaluate the ability of increasing concentrations of morphine to inhibit stimulated adenylyl cyclase activity in the presence of either 0.5 μM deltorphin II or 0.1 μM naltrindole. Ethanol extraction and analysis of samples by radioimmunoassay were performed following the manufacturer's instructions.
The activity of adenylyl cyclase was expressed as pmol of cAMP produced per mg protein in 5 min. The results were expressed as means ± SEM or as a percentage of the adenylyl cyclase activity stimulated by 50 μM forskolin. The concentration–response curves for cAMP were fitted to the logit–log equation. The morphine concentration–response curves in the absence and presence of δ-opioid receptor ligands were compared using two-way anova, with treatment as the between-groups factor and concentration as the repeated-measures factor. The level of significance was set at P < 0.05.
Co-immunoprecipitation experiments were performed as described by Garzón et al. (2005). Briefly, synaptosomal membranes from the whole brain (minus cerebellum) were prepared from both strains. To obtain synaptosomal extract, synaptosomes (600 μg of protein) were sonicated (two cycles of 5 s each) in 1 mL solubilization buffer containing 50 mM Tris–HCl (pH 7.7), 50 mM Na Cl, 1% Nonidet P-40, 50 μL of protease and phosphatase inhibitor mixtures (Sigma-Aldrich), and solubilized overnight at 4°C. The target proteins were immunoprecipitated from the synaptosomal extract with selected affinity purified biotinylated IgGs. The immunocomplexes were recovered with streptavidin agarose and the agarose pellets were solubilized in 2× Laemmli buffer. Biotinylated antibodies used were raised against: residues 208–216 of the second external loop of μ-opioid receptors (2EL; Garzón et al., 2005; Rodríguez-Muñoz et al., 2008); residues 110–120 of the first external loop of δ-opioid receptors (D1EL; Schulz et al., 1998; Garzón et al., 2005).
In some experiments, synaptosomal extracts were immunodepleted of μ receptors or δ receptors before the immunoprecipitation with biotinylated IgGs directed against the target proteins. To immunodeplete μ receptors, the synaptosomal extracts were subjected to an overnight incubation with the biotinylated antibody 2EL. Then, μ receptors were removed by incubating the extract with streptavidin agarose. The sample was centrifuged, and the supernatant was collected. Next, the supernatant was incubated overnight with the biotinylated antibody D1EL. The immunocomplexes were recovered with streptavidin agarose and the agarose pellets were solubilized in 2× Laemmli buffer. To immunodeplete δ receptors, the synaptosomal extracts were subjected to an overnight incubation with biotinylated antibody D1EL. Then, δ receptors were removed by incubating the extract with streptavidin agarose. The sample was centrifuged and the supernatant was collected. Next, the supernatant was incubated overnight with the biotinylated antibody 2EL. The immunocomplexes were recovered with streptavidin agarose, and the agarose pellets were solubilized in 2× Laemmli buffer.
Detection of immunoprecipitated proteins by immunoblotting
The recovered immunocomplexes were resolved on 10% SDS-PAGE and subsequently transferred onto 0.2 µm PVDF membranes (Bio-Rad 162–0176, Madrid, Spain) and probed with previously characterized antibodies directed towards: residues 2–16 of the extracellular N-terminal of the μ-opioid receptor (NT; Garzón et al., 2005); residues 2–16 of the extracellular N-terminal of the δ- opioid receptor (DNT; Garzón et al., 2005). The primary antibodies were detected using the corresponding secondary antibodies conjugated to horseradish peroxidase (diluted 1:10 000). Antibody binding was visualized with ECL Plus Western Blotting detection system (GE #RPN2132), and the chemiluminescence will be recorded with a ChemiImager IS-5500 (Alpha Innotech, San Leandro, CA) equipped with a Peltier-cooled CCD camera that provides a real-time readout of 30 frames per second. Data analysis was performed using the Image Gauge 4.0 software (Fujifilm Co., Barcelona, Spain).
All compounds for s.c. and i.p. injection were dissolved in normal saline, and injection volumes ranged from 0.2 to 0.6 mL. Morphine HCl was obtained from Alkaliber Laboratory (Madrid, Spain), methadone HCl from Dr Esteve Laboratories (Barcelona, Spain), fentanyl citrate from Roche Laboratories (Madrid, Spain), SNC80 and deltorphin II from Tocris (Bristol, UK), naltrindole HCl from Sigma-Aldrich (Madrid, Spain), and 3-d was provided by Dr Martín-Fontelles (University Juan Carlos I, Madrid, Spain). All other chemicals were of reagent grade and were obtained from Sigma-Aldrich. Radiochemicals and the cAMP assay kit were purchased from Tocris (Madrid, Spain), PerkinElmer (Madrid, Spain) or GE Healthcare (Barcelona, Spain).
Anti-nociceptive response to μ-opioid receptor agonists
The baseline responses before the administration of drugs were similar in both strains (tail flick test: Wistar = 3.8 ± 0.1 s, n= 51; SDU = 3.9 ± 0.1, n= 68; hot plate test: Wistar = 7.9 ± 0.3 s, n= 29; SDU = 7.2 ± 0.3 s, n= 28). The potency of morphine was 1.5–2-fold higher in SDU than in Wistar rats in both the tail flick and the hotplate tests (Figure 1A, Table 1). Statistical analyses of the dose–response curves obtained with the tail flick and hot plate tests indicated that strain had a significant effect (Table 1).
Table 1. Estimated ED50 values of μ-opioid receptor agonists in Wistar and SDU rats
We examined whether the higher sensitivity of SDU rats to morphine was specific to this opioid or was attributable to μ-opioid receptor-induced anti-nociception. Dose–response curves for the μ receptor agonists, fentanyl and methadone, and the highly selective μ receptor agonist, 3-d (Jagerovic et al., 2002) were obtained in the tail flick test. As before, the baseline response latency before drug administration was similar in SDU and Wistar rats. For each drug tested, the ED50 value was significantly lower in SDU than in Wistar rats (Table 1).
Anti-nociceptive response to the δ-opioid receptor agonist, SNC80
We studied whether the higher sensitivity of the SDU strain to μ receptor-induced anti-nociception was also observed with δ receptor-induced anti-nociception. Dose–response curves for the non-peptidic δ receptor agonist SNC80 in the tail flick test were obtained (Figure 1B), but we could not use doses of this agonist higher than 13.5 mg·kg–1 (%MPE: 40%, in both strains) because they induced restlessness. At the doses tested, no significant difference in the anti-nociceptive effect was apparent between strains.
Interaction of δ receptor antagonist naltrindole with the μ receptor agonists morphine and fentanyl
To test if interactions between δ- and μ-opioid receptors could account for the higher sensitivity of SDU rats to μ receptor agonists, dose–response curves for the μ receptor agonists morphine and fentanyl in the presence of the δ receptor antagonist naltrindole were obtained. Figure 1C shows that, in the tail flick test, subcutaneous administration of 10 mg·kg–1 naltrindole abolished the increased significant anti-nociceptive effect of 1.2 mg·kg–1 of morphine in SDU rats but not in Wistar rats (P < 0.05). This resulted in suppression of the significant differences in the estimates of ED50 of morphine and in the dose–response curves (Table 1). Naltrindole (10 mg·kg–1, s.c.) also reversed the increased effect of 10 and 20 μg·kg–1 of the highly selective μ receptor agonist fentanyl in SDU rats, compared with Wistar rats (P < 0.01) (Figure 1D).
Binding of [3H]naltrindole and [3H]DAMGO to the brain and spinal cord microsomal fraction
To determine the density of brain μ- and δ-opioid receptors in SDU and Wistar rats, saturation binding of [3H]naltrindole and [3H] DAMGO to the brain and spinal cord microsomal fraction was determined The apparent Kd values of the μ receptor agonist [3H]DAMGO and the δ receptor antagonist [3H]naltrindole in brain microsomal fractions were similar in SDU and Wistar rats, as was the number of μ and δ receptor binding sites (Figure 2A.B; Table 2). Inhibition of the binding of [3H]DAMGO and [3H]naltrindole demonstrated that the μ/δ receptor selectivity of CTAP was similar in brain membranes from either strain (>700-fold). The δ/μ receptor selectivity of unlabelled naltrindole was also similar in brain membranes from either strain (∼60-fold) (Figure 2C,D, Table 2). The inhibition of the binding of [3H] DAMGO by increasing concentrations of morphine and DAMGO revealed that both μ receptor agonists inhibited the binding of [3H] DAMGO with high affinity in rat brain membranes from both strains, according to a one-binding site model (Figure 3A,B, Table 2). In spinal cord, DAMGO inhibited the binding of [3H] DAMGO also with high affinity in both strains according to a one-binding site model (Wistar: Ki= 0.94 ± nM; SDU: Ki= 1.18 ± nM) (Figure 3C). Experiments on the inhibition of the binding of [3H]naltrindole to brain membranes by the μ receptor agonists morphine and DAMGO disclosed a clear difference between SDU and Wistar rats (Figure 3A,B). In the brain membranes of SDU rats, both μ receptor agonists inhibited the binding of [3H]naltrindole according to a two-binding site model (P < 0.01 in both cases). The percentage of high-affinity binding sites for morphine and DAMGO showed reasonable agreement. In contrast, for the brain membranes of Wistar rats, the results fitted significantly to a one binding-site model. The KI of morphine and DAMGO in the brain membranes of Wistar rats were similar to those for the low affinity sites in SDU rats (Table 2). In spinal cord membranes of both strains, DAMGO inhibited the binding of [3H]naltrindole according to a one binding-site model. The affinity of DAMGO was similar in both strains (Wistar: KI= 870 ± 127 nM; SDU: KI= 588 ± 74 nM) (Figure 3C).
Table 2. Affinities and number of binding sites for [3H]DAMGO and [3H]naltrindole binding to brain membranes from Wistar and SDU rats
Binding was performed as described in Methods. Values shown represent mean ± SEM of three experiments conducted in duplicate. Kd and KI values are expressed in nM. Bmax values are expressed in fmol·mg–1 protein. %1 represents the percentage of high-affinity binding sites.
0.58 ± 0.07
0.65 ± 0.06
133 ± 4
138 ± 3
2.09 ± 0.12
2.92 ± 0.37
0.84 ± 0.03
0.92 ± 0.03
7.10 ± 0.48
7.31 ± 0.92
8.95 ± 1.06
7.46 ± 0.73
0.031 ± 0.007
0.034 ± 0.008
75 ± 4
73 ± 3
19.94 ± 5.98
34.3 ± 0.13
238.5 ± 22.84
558.2 ± 101.5
19.64 ± 5.96
19.1 ± 0.3
946.5 ± 70.99
1466 ± 101.5
0.15 ± 0.02
0.12 ± 0.01
Binding of [35S]GTP-γ–S stimulated by μ- and δ-opioid receptor agonists
To determine whether the strains display differences in the effects of δ and μ receptor ligands on the activation of G-proteins, the binding of [35S]GTP-γ–S induced by μ and δ receptor agonists and the interactions between ligands for these receptors were assayed. As shown in Figure 4 and Table 3, no significant differences between strains were observed in the concentration–response curves of the μ or δ receptor agonists in brain and spinal cord membranes. However, for brain membranes, differences became obvious when concentration–response curves of δ receptor ligands were performed in the presence of a fixed concentration of μ receptor agonists. In brain membranes of Wistar rats, the presence of either 30 nM DAMGO or 30 nM morphine did not alter the binding of [35S]GTP-γ–S stimulated by increasing concentrations of the δ receptor agonist deltorphin II (Figure 5A, Tables 3 and 4), but in SDU brain membranes, the presence of either 30 nM DAMGO or 30 nM morphine increased the potency of deltorphin II by 5- to 10-fold (Figure 5B, Tables 3 and 4). In presence of morphine, the Emax also increased. In brain membranes, naltrindole had no significant effect on the binding of [35S]GTP-γ–S alone or in the presence of either 30 nM DAMGO or 30 nM morphine (data not shown). However, naltrindole inhibited the binding of [35S]GTP-γ–S stimulated by 3 μM DAMGO or 5 μM morphine in a concentration-dependent manner. In the case of Wistar rats, the concentration–response data fitted significantly to a one binding-site model. In SDU rats, experimental data fitted significantly to a two binding sites model. High-affinity phase represented approximately 45% of the binding (Figure 6A,B, Table 4). In the brain membranes of Wistar rats, the ID50 values for naltrindole were fairly similar to the ID50 of low affinity sites in SDU rats. The δ receptor antagonist TIPPΨ, did not affect the binding of [35S]GTP-γ–S to the brain membranes of either strain (Figure 7A,B). However, both 30 nM DAMGO or 30 nM morphine enhanced the binding of [35S]GTP-γ–S to SDU brain membranes in a concentration-dependent manner (Figure 7B). This reached a maximum at concentrations of approximately 10–7 M. At higher concentrations, a second inhibitory phase was also apparent. For the brain membranes of Wistar rats, in the presence of either 30 nM DAMGO or 30 nM morphine, TIPPΨ did not enhance the binding of [35S]GTP-γ–S (Figure 7A).
Table 3. Affinities and Emax of the stimulation of [35S]GTP-g-S binding to brain and spinal cord membranes from Wistar and SDU rats by μ− or δ-opioid receptor agonists
Binding was performed as described in Methods. Values shown represent mean ± SEM of three experiments conducted in triplicate. ED50 values are expressed in μM. Emax values are expressed as the percentage of basal binding.
0.08 ± 0.01
0.10 ± 0.01
135.0 ± 2.5
138.6 ± 2.7
0.15 ± 0.03
0.16 ± 0.05
130.1 ± 1.7
131.5 ± 2.6
0.25 ± 0.03
0.29 ± 0.03
129.9 ± 1.1
133.8 ± 1.4
0.07 ± 0.01
0.04 ± 0.01
114.5 ± 1.4
113.9 ± 0.9
0.83 ± 0.18
0.47 ± 0.05
114.0 ± 0.6
114.8 ± 0.3
0.14 ± 0.04
0.16 ± 0.04
127.9 ± 2.2
127.4 ± 1.8
Table 4. Affinities and Emax of the concentration-response curves of the [35S]GTP-g-S binding to brain membranes from Wistar and SDU rats by δ-opioid receptor agonists and antagonists in the presence of fixed concentrations of μ−opioid receptor agonists
Imax. Binding was performed as described in Methods. Values shown represent mean ± SEM of three experiments conducted in triplicate. ED50 values are expressed in μM. Emax and Imax values are expressed as the percentage of binding in the absence of δ receptor ligand.
In spinal cord membranes from both strains, 30 nM DAMGO did not have any significant effect on the ED50 of the binding of [35S]GTP-γ–S stimulated by increasing concentrations of deltorphin II (Wistar: ED50= 0.17 ± 0.05 μM; SDU: ED50= 0.17 ± 0.06 μM) (Figure 5C,D; Table 3). In spinal cord membranes, naltrindole had no significant effect on the binding of [35S]GTP-γ–S alone or in the presence of 30 nM DAMGO (data not shown). However, naltrindole inhibited the binding of [35S]GTP-γ–S stimulated by 3 μM DAMGO in a concentration-dependent manner (Figure 6C). For both strains, the concentration-response data fitted significantly to a one binding-site model. The ID50 values were fairly similar (Wistar: 0.19 ± 0.06 μM; SDU: 0.26 ± 0.09 μM). The δ receptor antagonist TIPPΨ, did not change binding of [35S]GTP-γ–S to spinal cord membranes of either strain, alone or in the presence of 30 nM DAMGO (Figure 7C,D).
Inhibition of adenylyl cyclase activity stimulated by forskolin
To determine whether there were differences between the strains in the inhibition of adenylyl cyclase activity, concentration–response curves of μ and δ receptor agonists and the interactions between μ receptor agonists and δ receptor ligands were carried out. The basal level of cAMP formation in the brain membranes of both strains (37.5 ± 5.2 and 28.3 ± 3.1 pmol·mg–1 protein·min–1; Wistar and SDU respectively) or that stimulated by 50 μM forskolin (234 ± 17 and 270 ± 30% stimulation; Wistar and SDU, respectively, n= 8) were not different. The parameters of the inhibition by either morphine or deltorphin II, of cAMP production stimulated by forskolin were the same in brain membranes from Wistar and SDU rats (Figure 8A,B, Table 5). However, the interaction of fixed concentrations of either deltorphin II or naltrindole with increasing concentrations of morphine was clearly strain–dependent. In the brain membranes from Wistar rats, both the potency and the Emax of morphine were not affected by the presence of 0.5 μM deltorphin II. In the presence of 0.1 μM naltrindole, the Emax of morphine was lower, whereas its potency was unchanged in Wistar rats (Figure 8C, Table 5). Analysis of the morphine concentration–response curves indicated a main effect of naltrindole (F1,24= 7.97; P < 0.01). In the brain membranes of SDU rats, deltorphin II and naltrindole shifted the concentration–response curves of morphine significantly to the left and right respectively (Figure 8D). In the presence of 0.5 μM deltorphin II (a concentration that by itself did not inhibit significantly the adenylyl cyclase activity stimulated by forskolin), the potency of morphine in inhibiting the adenylyl cyclase activity stimulated by forskolin was increased by approximately fivefold, with no change in the Emax (Table 5). In the presence of 0.1 μM naltrindole, morphine, at concentrations of up to 10–4 M, did not significantly inhibit the adenylyl cyclase activity stimulated by forskolin. Morphine only induced inhibition, of about 50%, at a concentration of 10–3 M. Analysis of the morphine concentration–response curves indicated a main effect of deltorphin II (F1,24= 4.32; P < 0.05) and naltrindole (F1,28= 19.11; P < 0.001) in SDU rats.
Table 5. Parameters of the concentration–response curves for morphine and deltorphin II, and morphine in the presence of deltorphin II or naltrindole in the inhibition of cAMP production in brain membranes from Wistar and SDU rats
n.a., not applicable. The activity of adenylyl cyclase was determined as described in Methods. Values shown represent mean ± SEM of three experiments conducted in triplicate. ID50 values are expressed in nM. Emax values are expressed as the percentage of the adenylyl cyclase activity stimulated by 50 μM forskolin.
498 ± 82
493 ± 98
60.3 ± 7.5
55.9 ± 5.0
847 ± 189
1920 ± 392
48.6 ± 4.2
52.9 ± 5.6
Morphine + Deltorphin II
983 ± 247
94.2 ± 15
51.3 ± 4.5
52.1 ± 2.8
Morphine + Naltrindole
382 ± 50
31.6 ± 1.2
Synaptosomal μ-opioid receptors were immunoprecipitated with a specific biotinylated anti-μ-opioid receptor antibody. When the immunoprecipitate was probed with a different anti-μ receptor antibody, a major band of ∼60 kDa was apparent. Minor bands of MWs ranging from 47 to 250 kDa were also evident. The 47 kDa corresponds to non-glycosylated μ receptors (Rodríguez-Muñoz et al., 2011). The higher MW bands represent different degrees of glycosylation (Garzón et al., 1995; 2005). Only the major band was analysed. No differences in the amount of this band between strains were observed (Figure 9A,B). The antibody used precipitated almost all μ receptors, as a second immunoprecipitation of the unbound synaptosomal extract revealed practically no μ receptors (Figure 9C,D). When immunoprecipitated synaptosomal μ receptors were probed with an anti-δ receptor antibody, a major band of 58 kDa was observed. Minor bands ranging from 50 to 165 kDa were also evident. Given that it has been demonstrated that the high MW bands represent different degrees of glycosylation (Garzón et al., 2005), only the major band was analysed. The amount of the δ receptor 58 kDa band associated with immunoprecipitated μ receptor was significantly higher in SDU than in Wistar rats (Figure 9E,F). Immunodepletion of δ receptors before the immunoprecipitation of μ receptors reduced the intensity of the major band of μ receptors in both strains. A larger reduction was evident in SDU compared with Wistar, but this difference was not statistically significant (Figure 9G,H).
To confirm the results of immunoprecipitation with the μ receptor antibody, a parallel set of experiments immunoprecipitating δ receptors with a specific biotinylated δ receptor antibody was performed. When the immunoprecipitated δ receptors were probed with a different anti-δ receptor antibody, a major band of 58 kDa and minor bands ranging from 37 to 165 kDa were observed. As before, only the major 58 kDa band was analyzed and there were no differences between strains in the extent of this δ- receptor band (Figure 10A,B). The antibody used precipitated almost all δ receptors, as a second immunoprecipitation of the unbound synaptosomal extract revealed practically no δ receptors (Figure 10C,D). In agreenment with the earlier experiments, the extent of the μ receptor 60 kDa band associated with immunoprecipitated δ receptors was significantly higher in SDU than in Wistar rats (Figure 10E,F). Immunodepletion of μ receptors before the immunoprecipitation of δ receptors resulted in a significantly higher reduction of the 58 kDa band δ receptors in SDU than in Wistar rats (Figure 10G,H).
The most significant finding presented here is that δ-opioid receptors determine the sensitivity to the anti-nociceptive effects of μ-opioid receptor agonists. Using different experimental approaches, we have provided evidence that the extent of the interactions between μ and δ-opioid receptors varies between two strains of rats that show different sensitivities to the anti-nociceptive effect of μ receptor agonists.
We have demonstrated an enhanced anti-nociceptive potency of agonists of μ receptors, but not δ receptors, in SDU rats when compared with Wistar rats, in nociceptive tests. It could be argued that the excitatory effects of SNC80 (Jutkiewicz et al., 2004) could bias the apparent anti-nociception. Nonetheless, the fact that, at all the doses of SNC80 tested, similar results were obtained in both strains strongly suggests that the anti-nociceptive effect induced by δ receptor agonists did not differ between SDU and Wistar rats. Differences in the anti-nociceptive effect induced by morphine or fentanyl were already evident at low doses (0.6 mg·kg–1 and 10 μg·kg–1 respectively). Hence, SDU rats displayed enhanced anti-nociception that was mediated by μ receptors. The higher sensitivity to morphine and the highly selective μ receptor agonist fentanyl, shown by SDU rats depends on δ receptor modulation of the μ receptor-mediated anti-nociception because, in the presence of the selective δ receptor antagonist naltrindole, the anti-nociception induced by morphine or fentanyl in SDU rats was restored to values that resembled those in Wistar rats. Naltrindole did not alter the anti-nociception induced by morphine and fentanyl in Wistar rats. The dose of naltrindole used here is selective for δ-opioid receptor antagonism (Portoghese et al., 1988; Ossipov et al., 1994), although at higher doses, naltrindole has agonist effects at δ- and κ-opioid receptors (Stapelfeld et al., 1992; Takemori et al., 1992; Stewart and Hammond, 1994). Therefore, our data suggest that a mechanism that involves activation of δ receptors underlies the enhanced sensitivity of SDU rats to anti-nociception induced by μ receptor agonists. Several reports have suggested that anti-nociceptive activity mediated by μ receptors can be modulated by δ receptors. In the present study, the modulation observed is positive. This has also been described by others (Lee et al., 1980; Vaught et al., 1982; Ward and Takemori, 1983; Heyman et al., 1989; Negri et al., 1995; Gomes et al., 2004). Other reports suggest a negative regulation (Lee et al., 1980; Heyman et al., 1989; Schiller et al., 1999; Gomes et al., 2004; Guan et al., 2005; He et al., 2011). Many factors could account for these apparently conflicting results, such as the ligands, route of administration, level of the interaction, animal species or experimental approach.
Interactions between δ- and μ-opioid receptors were evident in radioligand-binding, stimulation of [35S]GTP-γ–S binding and inhibition of adenylyl cyclase activity in brain membranes of SDU but not Wistar rats. Labelled antagonists are useful to detect changes in the binding properties of agonists. They have the same affinity towards different states of receptors and allow the detection of any change in affinity of competing agonists (Georgoussi et al., 1993; Garzón et al., 1998). In Wistar brain membranes, μ receptor agonists inhibited the specific binding of the δ receptor antagonist [3H] naltrindole with low affinity. In SDU, an additional high-affinity binding site was apparent. In spinal cord, only one low affinity binding site was detected in both strains. The δ/μ receptor selectivity of naltrindole and the μ/δ receptor selectivity of CTAP are nearly identical in the brain membranes of both strains. Thus, the possible binding of [3H] naltrindole to μ receptors, inhibited by μ receptor agonists in the brain of SDU rats, but not Wistar, can be ruled out.
In the brains of SDU, but not Wistar, rats the binding of [35S]GTP-γ-S induced by the δ receptor agonist deltorphin II and the δ receptor antagonist TIPPΨ was enhanced in the presence of low concentrations of μ receptor agonists, suggesting a cross-regulation between μ and δ receptors in SDU rats. Naltrindole inhibited the binding of [35S]GTP-γ-S induced by high concentrations of μ receptor agonists with low affinity in Wistar brain membranes. In SDU brain membranes, an additional high-affinity binding site was apparent. In SDU brains, the ratio of the IC50s of the low and high-affinity phases was ∼170. This ratio is similar to the δ/μ receptor selectivity ratio of naltrindole for the inhibition of GTP-γ–35S binding to δ and μ receptors cloned in COS cells (143-fold) (Xu et al., 2001). In contrast to a previous study, the results obtained in spinal cord membranes did not show any significant interaction in the binding of [35S]GTP-γ–S induced by δ receptor ligands in the presence of a μ receptor agonist (Gomes et al., 2004). The use of different animal species may explain this discrepancy.
We also examined the interaction of δ and μ receptors on the inhibition of forskolin-stimulated adenylyl cyclase activity by morphine. The δ receptor agonist deltorphin II increased the potency of morphine only in SDU rats. The δ receptor antagonist naltrindole decreased the Emax of morphine in Wistar rats. In contrast, in SDU rats in the presence of 100 nM naltrindole morphine, at concentrations up to 100 μM, did not have any effect on forskolin-stimulated cAMP production. It has been reported that 100 nM naltrindole attenuates the inhibition of forskolin-stimulated cAMP production induced by 1 μM DAMGO only in some brain areas (Izenwasser et al., 1993). Thus, the effect of 100 nM naltrindole in Wistar rats was probably due to its binding to μ receptors. In SDU rats, additional mechanisms, probably through δ receptors, could be involved. These results indicate that interactions between δ and μ receptors in vivo modifies signal transduction in our model. Taken together, these experiments, as well as the radioligand and [35S]GTP-γ-S binding experiments, provide a molecular basis for the greater anti-nociceptive effect of μ receptor agonists in SDU rats and its reversal by a δ receptor antagonist.
Classically, it was thought that a particular effect mediated by a drug or other ligand resulted from its interaction with a single type of receptor. However, there is increasing evidence that interactions between receptors play a major role in the actions of opioids (Smith and Lee, 2003). Since the late 1970s, compelling data from functional and radioligand binding studies have suggested that μ and δ receptors can interact (Vaught and Takemori, 1979). There is evidence for modulation of μ receptor-mediated analgesia by δ receptor agonists (Vaught et al., 1982), allosteric coupling between μ and δ receptors (Rothman et al., 1993), co-localization of opioid receptors in some neurons (Ji et al., 1995) and the possible existence of a μ-δ receptor complex (Schoffelmeer et al., 1990; Gomes et al., 2004; Garzón et al., 2005). The functional role of these interactions upon co-expression of both opioid receptors is also evident in vitro (Snook et al., 2008), as well as in mice lacking the μ receptor (Matthes et al., 1998) or the δ receptor (Zhu et al., 1999; Scherrer et al., 2004). In mice lacking δ-opioid receptors, δ receptor agonists still produce supraspinal analgesia, suggesting the existence of a second δ receptor-like analgesic system (Zhu et al., 1999) or μ receptor-dependent anti-nociception (Scherrer et al., 2004). In our study, the different effect of naltrindole on anti-nociception induced by μ receptor agonists in SDU and Wistar rats cannot be attributed to the binding of naltrindole to μ receptors in the SDU strain. The μ and δ receptors can coexist in a complex with a series of signalling elements such as G-proteins and RGS proteins (Garzón et al., 2005). We found, in co-immunoprecipitation experiments, that the μ receptors were associated with δ receptors to a greater extent in SDU than in Wistar rats. Whether such association results from direct specific interactions, is promoted indirectly by scaffolding proteins or results from the high density of receptors within given lipid microdomains remain to be established (Snook et al., 2006).
It has been reported that μ-δ-opioid receptor heteromers generate distinct pharmacological properties, such as the enhancement of the μ receptor binding and signalling by δ receptor antagonists (Gomes et al., 2004). We detected some association of μ receptors with δ receptors in brains of Wistar rats, yet no functional or pharmacological evidence was observed in anti-nociception, radioligand binding or signal transduction. Thus, the differences between Wistar and SDU rats cannot be explained simply by the absence of μ-δ-receptor heteromers in Wistar and their presence in SDU rats. In the CNS, opioid receptors exist in a complex with G-proteins, RGS proteins, the HINT1 protein and scaffolding proteins. The affinity of agonists to μ-opioid receptors depends on the nature of G-protein coupled to the μ receptor (Garzón et al., 1998). RGS proteins regulate the effect of opioid receptor agonists. For instance, knockdown of the RGS7 protein, a member of the R7 subfamily of RGS proteins, enhances the anti-nociceptive potency of μ- and δ-opioid receptor agonists (Garzón et al., 2003). HINT1 forms homodimers in cells and these dimers act in the cross-regulation between NMDA receptors and μ-opioid receptors (Rodríguez-Muñoz et al., 2011). The possible interaction of different signalling proteins with μ receptors could underlie the increased sensitivity to μ receptor agonists in SDU rats. This interaction and its regulation by δ receptors should be further explored.
In conclusion, we report here an in vivo animal model in which the degree of μ- and δ-opioid receptor cross-regulation at supraspinal level was correlated with the sensitivity to the anti-nociceptive effect of μ-opioid receptor agonists. In vivo interactions between exogenous μ-opioid receptor agonists and endogenous δ-opioid receptor peptides may produce the differential pharmacological effects reported here in the nociceptive tests. The lack of correlation between the in vivo sensitivity to opioids and baseline results of the in vitro assays could be explained by the absence of endogenous peptides in the in vitro assays. Provided that the results presented here are relevant to the clinical setting, one future challenge is to determine which factors influence the degree of μ- and δ-opioid receptor cross-regulation. This strategy, together with the identification of human biomarkers for opioid sensitivity, could represent the next step in pain relief.
We thank B Fraile and E Sabater for technical assistance, Dr MI Martin-Fontelles for supplying 3-d and Dr LA Parker for English revision. This work was supported by grants from the Spanish MCyT (SAF 2000-0120-C02-02 to CCF and SAF 2003-08140-C02-02 to JJB), FIS (FIS05/PI050590 to CCF) and Instituto de Salud Carlos III (PS09/00332 to JG). JC was a recipient of predoctoral fellowships from the Instituto Upsa de Dolor and Generalitat Valenciana-Fondo Social Europeo.