Agonist Activity of the δ-Antagonists TIPP and TIPP-ψ in Cellular Models Expressing Endogenous or Transfected δ-Opioid Receptors

Experimental Procedures

Materials.

Penicillin/streptomycin (10,000 IU/ml and 10,000 μg/ml), Geneticin (G418), HAT (0.1 mM hypoxanthine, 10 μM aminopterin, and 17 μM thymidine; 50X), fetal calf serum, and DMEM containing 4.5 g glucose, l-glutamine, and pyruvate were purchased from Mediatech Cellgro (Herndon, VA). Hygromycin-B was supplied by Calbiochem (San Diego, CA). OptiMEM and the transfection reagent Lipofectin were purchased from Invitrogen (Rockville, MD). The pREP4 plasmid was purchased from Invitrogen (San Diego, CA). Naloxone and DPDPE were obtained from Peninsula Laboratories (Belmont, CA) and TIPP from Phoenix Pharmaceuticals, Inc. (Belmont, CA). TIPP-ψ was a generous gift from Dr. Tim Hales (George Washington University, Washington, D.C.). [³H]Diprenorphine (56 Ci/mmol) was purchased from PerkinElmer Life Science Products (Boston, MA), and [8-³H]adenine (26 Ci/mmol) was purchased from Amersham Pharmacia Biotech (Piscataway, NJ). All other reagents were purchased from Fisher Scientific (Pittsburgh, PA).

Transfection.

GH₃ cells (CCL 82.1) were stably transfected with cDNA encoding for μ-opioid (GH₃MOR) or both μ- and δ-opioid (GH₃MORDOR) receptors as previously described (Piros et al., 1995; Prather et al., 2000). CHO and HEK cells transfected with cDNA encoding for δ-opioid receptors (CHO-DOR and HEK-DOR, respectively) were provided by Dr. Tim Hales (George Washington University, Washington, DC). To produce the GH₃DORT cell lines, GH₃cells were stably transfected with pREP4 plasmids containing cDNA encoding for δ-opioid receptors with the hemagglutinin epitope tag spliced at the N terminus (DORTAG) (Ko et al., 1999). Dr. Ping-Yee Law (University of Minnesota, Minneapolis, MN) generously provided the DORTAG/pREP4 construct. Specifically, cells were seeded into 100-mm dishes and cultured until approximately 40% confluency. Following a single wash with serum-free OptiMEM, a mixture of 50 μg of Lipofectin reagent and 10 μg of the DORTAG/pREP4 construct (i.e., a 5:1 ratio) in 7 ml of serum-free OptiMEM was added. Cells were incubated under these conditions for 6 h at 37°C in 5% CO₂/95% air. The transfection media was replaced with 20 ml of normal growth media without selection antibiotic for 2 days. After splitting each 100-mm dish into two 150-mm dishes, selection of GH₃ cells that stably incorporated the DORTAG/pREP4 plasmids was initiated by the inclusion of 500 μg/ml of hygromycin-B in the growth media. Colonies that survived culturing in the presence of the selection antibiotic were picked and cultured until a sufficient number of cells were obtained for screening. Confirmation of δ-opioid receptor expression was determined by performing competition for [³H]diprenorphine (0.5 nM) binding by naloxone (10 μM) as described below. Forty-eight clones that demonstrated a wide range of specific [³H]diprenorphine binding were selected; of these, five were selected for future studies in which saturation binding was performed to determine the actualK_d andB_max for [³H]diprenorphine.

Cell Culture.

All cells were maintained in a DMEM-based media supplemented with NaHCO₃ (3.7 g/l), 10% (v/v) fetal calf serum, 100 units/ml penicillin, and 100 μg/ml streptomycin. GH₃MORDOR transfected cells used this growth media supplemented with 2.5 mg/ml geneticin (G418) and 200 μg/ml hygromycin-B, and GH₃DORT cells supplemented this growth media with 200 μg/ml hygromycin-B. CHO-DOR and HEK-DOR cells were maintained in the growth media supplemented with 2.5 mg/ml geneticin (G418). NG108 cells were maintained in growth media supplemented with HAT. No additional supplements were required for wild-type GH₃ or N1E115 cells. Cells were incubated in a humidified atmosphere of 5% CO₂/95% air at 37°C and harvested with a 10 mM phosphate-buffered saline solution containing EDTA (1 mM), pH 7.4. Cells harvested for adenylyl cyclase assays were re-seeded at a density of 8 × 10⁶ cells per 17-mm (24-well) culture plate. Cells collected for membrane preparation were centrifuged (1000 rpm, 4°C, 10 min) and the pellets stored at −80°C until used.

Membrane Preparation and Receptor Binding.

GH₃ membranes containing the transfected opioid receptor(s) of interest were prepared for binding assays as follows. Harvested cell pellets were thawed on ice, resuspended in ice-cold homogenization buffer [50 mM HEPES (pH 7.4), 3 mM MgCl₂, and 1 mM EGTA], followed by homogenization with 10 strokes using a glass Dounce homogenizer and pestle A (Wheaton, Philadelphia, PA). The cell homogenates were centrifuged at 18,000 rpm for 10 min at 4°C, the supernatant was discarded, and the resultant pellet resuspended in the original volume of homogenization buffer. The procedure was repeated twice more, and the partially purified membrane pellet was re-suspended in 50 mM Tris, pH 7.4, at 10% of the original volume. Protein concentration was determined by the Lowry method (Lowry et al., 1951), and aliquots were stored at −80°C.

Saturation and competition binding assays were performed in 50 mM Tris, pH 7.4, with 5 mM MgCl₂, at room temperature using a 90-min incubation period, as described previously (Prather et al., 2000). For saturation binding studies, 0.05 to 30 nM [³H]diprenorphine was used, with nonspecific binding determined by the presence of 5 μM naloxone. In competition binding experiments, the ability of increasing concentrations DPDPE (0.1 nM to 3 μM), TIPP, and TIPP-ψ (0.01 nM to 1 μM) to displace the binding of [³H]diprenorphine (1 nM) was assessed. Binding reactions were terminated by filtration with a Brandel 24-sample standard format cell harvester (Gaithersburg, MD), and after the addition of 4-ml scintillation fluid, the amount of radioactivity on the filters was determined using a Packard Tri-Carb 2100TR liquid scintillation counter (Meriden, CT).

Measurement of cAMP Levels.

The effect of opioids on the conversion of [³H]adenosine triphosphate (ATP) to cyclic [³H]adenosine monophosphate (cAMP) by adenylyl cyclase was determined as previously described (Law et al., 1983a). Briefly, cells were seeded into 24-well plates and cultured for 24 h (∼90% confluency). On the day of the assay, medium was removed and replaced with an incubation mixture (37°C) of DMEM containing 0.9% NaCl, 500 μM 3-isobutyl-1-methylxanthine and 1.25 μCi/well [³H]adenine for 2 h. After incubation, the mixture was removed, and each plate was floated in an ice-water bath for 5 min. During this time, an assay mixture of ice-cold Krebs-Ringer-HEPES buffer containing 500 μM 3-isobutyl-1-methylxanthine, 10 μM forskolin, and the appropriate concentration of the opioid ligand of interest was added. Due to limited solubility, TIPP was dissolved in a 1% DMSO solution, whereas TIPP-ψ and DPDPE were soluble in water. The presence of DMSO was controlled for in experiments containing TIPP. Plates were then floated in a water bath at 37°C for 15 min. The reaction was terminated by the addition of 50 μl of 2.2 N HCl. Radioactive cAMP was separated using alumina column chromatography (Alvarez and Daniels, 1992). Scintillation fluid (10 ml) was added to each sample prior to counting in a Packard Tri-Carb 2100TR liquid scintillation counter (Meriden, CT).

Isobologram Analysis.

Isobolographic analysis was performed using the method previous described (Tallarida et al., 1989; Martin and Prather, 2001). Briefly, the IC₅₀ value for a single ligand (i.e., DPDPE or TIPP) was determined from the adenylyl cyclase assay. The assay was then repeated with the ligands coadministered at a constant dose ratio based on an equieffective dose. Equieffective dose ratios were based on the IC₅₀for each ligand in the adenylyl cyclase assays. For example, if the IC₅₀ of ligand A was 20 nM and the IC₅₀ of ligand B was 5 nM, then the equieffective dose ratio of A to B was 4:1. Therefore, at each concentration of the concentration-effect curve, the concentration of drug A was always 4 times the concentration of drug B. The IC₅₀ value for each ligand in the presence of the other coadministered ligand was then determined by analyzing two separate curves in which the same Y values (i.e., % effect) were plotted against two different X values (i.e., concentrations for each ligand). The isobolograph was constructed by plotting the experimentally determined IC₅₀ values for DPDPE and TIPP administered alone on the X and Y axes, respectively. The diagonal, linear regression line connecting these values represents the theoretical line of additivity. On the additivity line lies the theoretical IC₅₀. The X/Y coordinates of the theoretical IC₅₀ are calculated for each ligand when coadministered in equieffective concentrations and represent the IC₅₀ values if the interactions were purely additive. The experimental IC₅₀ values of each ligand when coadministered then provided the X/Y coordinates for the observed IC₅₀ to be graphed on the isobolograph. If the combination of ligands resulted in only an additive interaction, the observed IC₅₀ was on the line of additivity. An observed IC₅₀ significantly below the line of additivity indicated a synergistic (or greater than additive) interaction between ligands. An observed IC₅₀significantly above the additivity line suggested a less than additive interaction between ligands.

Data Analysis and Statistics.

Determination of receptor affinity (K_d) and receptor density (B_max) for saturation binding experiments was performed using the nonlinear regression analysis of GraphPad Prism v3.0 (GraphPad Software, San Diego, CA). The IC₅₀ values from the competition binding experiments were also determined using GraphPad Prism. The conversion of IC₅₀ to K_ivalues was calculated using the Cheng-Prusoff equation (Cheng and Prusoff, 1973). Data are expressed as mean ± S.E.M. or mean (95% confidence interval), as indicated. Unless otherwise stated, data are represented by 3 separate experiments, done in duplicate or triplicate. Statistical significance of the data was determined by a one-way ANOVA, followed by comparison using the Tukey (comparison of all conditions) and Dunnett's (comparison to control) multiple comparison post-tests.

Results

Stimulation of opioid receptors by agonists results in the activation of G-proteins and the subsequent inhibition of adenylyl cyclase activity (Standifer and Pasternak, 1997). This inhibition is reflected by a decrease in the formation of intracellular cAMP. In contrast, binding of opioid receptor antagonists does not activate G-proteins, and therefore no inhibition of adenylyl cyclase activity occurs. In a previous study examining μ-/δ-opioid receptor interactions in GH₃ cells (Martin and Prather, 2001), the prototypical δ-opioid receptor antagonist TIPP was used as a pharmacological probe. Unexpectedly, TIPP displayed agonist activity in GH₃MORDOR cells when a maximally effective concentration (1 μM) of TIPP inhibited adenylyl cyclase activity by 52.2% (Fig. 1; Table1). This inhibition was slightly, but significantly, less than the inhibition of DPDPE (69.2%), a known δ-opioid receptor agonist. Interestingly TIPP-ψ (1 μM), a chemically and enzymatically stable derivative of TIPP, also significantly inhibited adenylyl cyclase (38.4%), although less than TIPP and DPDPE. To determine whether this apparent agonist activity was mediated through δ-opioid receptors or due to the coexpression of μ- with δ-opioid receptors, the activity of TIPP and TIPP-ψ was compared between GH₃DORT-8 (δ-receptor only), GH₃MOR (μ-receptor only) and wild-type GH₃ (no transfection) cells. In GH₃DORT-8 cells, TIPP inhibited adenylyl cyclase activity (69.7%) to an extent that did not differ significantly from the full δ-agonist, DPDPE (78%) (Fig. 1; Table 1). Also, TIPP-ψ inhibited adenylyl cyclase activity (54%) in GH₃DORT-8 cells similar to its activity in GH₃MORDOR cells. No significant inhibition was produced by TIPP, TIPP-ψ, or the positive control, DPDPE, in GH₃MOR or wild-type GH₃cells. Competition binding studies with [³H]diprenorphine in GH₃DORT-8 membranes showed TIPP exhibited a high affinity for the δ-opioid receptor (K_i = 2.00 nM), whereas the affinity of TIPP-ψ was slightly less (K_i = 13.1 nM) (Table 2). These results indicate that TIPP and TIPP-ψ exert their effects through the δ-opioid receptor, exhibit high affinity for the δ-opioid receptor, and that in GH₃DORT-8 cells the inhibition by TIPP is not statistically different from the full δ-agonist, DPDPE.

Concentration-dependent effect and maximally effective concentrations of TIPP and TIPP-ψ were established and compared with DPDPE by measuring the inhibition of adenylyl cyclase activity with increasing concentrations of all three drugs (Fig.2, closed symbols). Results showed that TIPP and TIPP-ψ concentration-dependently inhibited adenylyl cyclase activity in GH₃DORT-8 cells. Half-maximal inhibition was achieved at concentrations of 0.162, 3.97, and 0.293 nM for TIPP, TIPP-ψ, and DPDPE, respectively (Table 2). Pretreatment of GH₃DORT-8 cells with PTX (100 ng/ml) for 24 h attenuated the maximal inhibition produced by all three ligands, indicating the involvement of G_iα/G_oα G-proteins (Fig. 2; open symbols). These results indicate that inhibition of adenylyl cyclase activity by TIPP and TIPP-ψ is receptor-mediated, dose-dependent, and involves the G_iα/G_oα subfamily of G-proteins. Of note was that maximal inhibition of adenylyl cyclase activity was achieved at concentrations of 10 nM or greater for both TIPP and DPDPE, and 100 nM or greater for TIPP-ψ. Due to limited quantities of the noncommercially available peptide TIPP-ψ, 100 nM was selected to produce maximal effects for all subsequent experiments, whereas 1 μM was selected for DPDPE and TIPP.

As the density of δ-opioid receptors in GH₃MORDOR (3.09 pmol/mg) and GH₃DORT-8 (2.16 pmol/mg) cells were quite high, relative to levels expressed in brain tissue (Table 1), the issue of receptor density versus effect was addressed. A series of clones were generated by the stable transfection of GH₃ cells with hemagglutinin tagged δ-opioid receptors (GH₃DORT) (Fig. 3). Selected clones contained a 28-fold range of δ-opioid receptor densities (0.08–2.25 pmol/mg), as determined by saturation binding with [³H]diprenorphine (Table 1; shown in parentheses, Fig. 3). TIPP and TIPP-ψ significantly inhibited adenylyl cyclase activity in all clones regardless of receptor density, as did the δ-agonist, DPDPE. As anticipated, the efficacy increased in direct proportion to receptor density, and DPDPE reached maximal efficacy of inhibition at a lower receptor density than TIPP or TIPP-ψ. Interestingly, the relative rank order of efficacy was DPDPE = TIPP > TIPP-ψ, with the exception of GH₃DORT-2, where DPDPE > TIPP > TIPP-ψ. These results demonstrate that maximal inhibition by TIPP and TIPP-ψ was significant in GH₃ cells expressing a wide range of δ-opioid receptor densities, including those physiologically relevant, and is independent of receptor density. In addition, these results demonstrate that the agonist activity of the δ-antagonist TIPP was equivalent to that of the full δ-agonist DPDPE in 3 of the 4 GH₃DORT clones tested.

As both TIPP and TIPP-ψ are selective δ-opioid receptor antagonists, we examined other selective δ-antagonists (i.e., naltrindole, naltriben, naloxone, and ICI 174864) for agonist activity in our paradigm using the GH₃DORT-8 clone (Fig.4). The GH₃DORT-8 clone was chosen for this, and all subsequent experiments, because it expressed the lowest δ-receptor density at which both TIPP and TIPP-ψ demonstrated maximal inhibition of adenylyl cyclase activity. TIPP (100 nM) produced 58% inhibition, and this level of inhibition did not differ significantly from that of the full δ-agonist, DPDPE (70.7%). TIPP-ψ, a derivative of TIPP, also produced significant inhibition (49.4%), although less than that observed for either TIPP or DPDPE. Naltrindole (1 μM) acted as a partial agonist, producing significant inhibition (35.1%), but less than that shown for DPDPE, TIPP, or TIPP-ψ. In contrast, naltriben and naloxone (1 μM) acted as neutral antagonists and did not produce significant inhibition (17% and 9.4%, respectively). ICI 174864, however, displayed significant inverse agonist activity, resulting in a 21.5% stimulation of adenylyl cyclase activity. These results importantly demonstrate that not all δ-opioid receptor antagonists inhibit adenylyl cyclase in GH₃DORT-8 cells, but rather show a range of efficacy varying from inhibition (TIPP), equivalent to that of the full δ-agonist DPDPE, to stimulation (ICI 174864) of adenylyl cyclase activity.

It was next determined if the apparent agonist activity of TIPP could be reversed by selective δ-antagonists. A maximally effective concentration of TIPP (10 nM) produced significant inhibition of adenylyl cyclase activity (65%) in GH₃DORT-8 cells (Fig. 5). Addition of the neutral antagonists naltriben and naloxone (10 μM) attenuated the inhibition produced by TIPP to 20 and 22%, respectively, but did not completely reverse the inhibition to levels of control. The δ-antagonist ICI 174864 (10 μM) not only completely reversed the inhibition by TIPP, but also produced significant stimulation of adenylyl cyclase activity (11.1%) relative to control. In comparison, a less than maximal concentration of DPDPE (0.5 nM) significantly inhibited adenylyl cyclase activity (48.2%) in GH₃DORT-8 cells. Addition of ICI 174864 attenuated inhibition by DPDPE to only 11.8%, but did not completely antagonize the inhibition to levels significantly different from control. Interestingly, the addition of TIPP (10 nM) to DPDPE (0.5 nM) increased, rather than antagonized, the inhibition of adenylyl cyclase activity. This suggested an additive effect, which was further examined by coadministration studies using full concentration-effect curves and isobolographic analysis (Fig.6). IC₅₀ values for the reduction of cAMP levels were determined for DPDPE and TIPP administered alone and were found to be 0.293 nM and 0.162 nM, respectively (see Fig. 2; Table 2). Upon coadministration of TIPP and DPDPE in equieffective concentrations, based on a 1:1.8 constant dose ratio (refer to Experimental Procedures), respectively, the determined IC₅₀ values of TIPP and DPDPE in the presence of the coadministered drug were 0.081 and 0.146 nM, respectively. These observed values did not differ significantly from the theoretical values (0.091 nM, 0.162 nM) (Fig. 6). Therefore, these results indicate that coadministered TIPP and DPDPE interact in an additive manner to inhibit adenylyl cyclase activity in GH₃DORT-8 cells.

The agonist activity of TIPP and TIPP-ψ was further examined in other cellular models containing endogenous and transfected δ-opioid receptors (Fig. 7). N1E115 (mouse neuroblastoma) and NG108-15 (mouse neuroblastoma × rat glioma) cells express endogenous δ-opioid receptors at densities of 0.120 and 0.570 pmol/mg, respectively, as determined by previous studies (Law et al., 1983b; Prather et al., 1994) (Table3). TIPP and TIPP-ψ significantly inhibited adenylyl cyclase activity in both N1E115 and NG108-15 cells. The relative rank order of efficacy was similar to the GH₃DORT clones in N1E115 cells, where DPDPE = TIPP > TIPP-ψ. In NG108-15 cells, the maximal inhibition produced by TIPP and DPDPE did not differ significantly; however, the inhibition did differ significantly between that shown for TIPP-ψ and DPDPE. Similarly, TIPP and TIPP-ψ significantly inhibited adenylyl cyclase in the transfected CHO-DOR and HEK-DOR cell lines (Fig. 7). The receptor densities of these transfected cell lines were determined in the present study, and ranged from very low (CHO-DOR; 0.151 pmol/mg) to very high (HEK-DOR; 6.69 pmol/mg) (Table 3). In CHO-DOR, the relative rank order of efficacy was DPDPE > TIPP = TIPP-ψ. In contrast, the relative rank order of efficacy in HEK-DOR cells was DPDPE = TIPP > TIPP-ψ. These results address two important issues. First, these results support the earlier finding in GH₃DORT cells that the agonist activity of TIPP and TIPP-ψ is not dependent on receptor density. Significant inhibition by TIPP and TIPP-ψ was seen in cells containing either endogenous or transfected δ-opioid receptors, over a 55-fold span of receptor densities. Second, the agonist activity of TIPP and TIPP-ψ is not limited to the δ-opioid receptor cDNA used to transfect GH₃ cells. That inhibition was observed in other cells containing either endogenous or transfected δ-opioid receptors supports this finding.

Discussion

A significant body of evidence indicates that TIPP and TIPP-ψ act as selective δ-opioid receptor antagonists, which demonstrate high affinity, potency, and selectivity (for reviews, see Schiller et al., 1999a,b). The antagonist profile of these compounds was primarily determined by using the mouse vas deferens assay, in which TIPP and TIPP-ψ failed to show any activity, but blocked the activity of known δ-opioid agonists. In agreement with these findings, other in vitro studies report that TIPP and TIPP-ψ do not stimulate GTPase activity or [³⁵S]GTPγS binding, indicating that these compounds do not activate G-proteins (Mullaney et al., 1996; Szekeres and Traynor, 1997). In sharp contrast, the current study demonstrates that TIPP and TIPP-ψ can act as δ-opioid receptor agonists, as reflected by their ability to inhibit adenylyl cyclase activity, and involve G_iα/G_oαG-proteins.

TIPP and TIPP-ψ were found to significantly inhibit adenylyl cyclase activity in GH₃DORT or GH₃MORDOR cells. Neither compound demonstrated any effect on adenylyl cyclase activity in wild-type GH₃ cells or GH₃MOR cells, thus indicating that the agonist activity required the presence of the δ-opioid receptor. In fact, this agonist activity proved to be a concentration-dependent, δ-receptor-mediated effect involving the G_iα/G_oα G-proteins. Pretreatment of GH₃DORT-8 cells with PTX (24 h, 100 ng/ml) attenuated, but did not completely block, the inhibition of adenylyl cyclase activity by TIPP, TIPP-ψ, and the δ-agonist, DPDPE. There are two possible explanations for the incomplete effect of PTX. The first, and simplest, explanation is that greater concentrations of PTX were needed to completely ADP-ribosylate all G_iα/G_oα proteins. The second explanation is that a portion of the δ-receptor-mediated inhibition of adenylyl cyclase involved PTX-insensitive G-proteins (i.e., G_qα). This second explanation seems unlikely because PTX pretreatment in GH₃MORDOR cells completely reversed the reduction of cAMP levels produced by TIPP and TIPP-ψ (data not shown). Regardless, the inhibition of all three ligands was reduced by an equivalent amount suggesting the involvement of the same, or similar, G-proteins.

Because both GH₃MORDOR and GH₃DORT-8 cells contain relatively high receptor densities, we tested both compounds at maximally effective concentrations in GH₃DORT clones containing a 28-fold range of δ-receptor densities (0.080–2.25 pmol/mg). TIPP and TIPP-ψ significantly inhibited adenylyl cyclase activity in all GH₃DORT clones, regardless of receptor density. These results indicated that the agonist activity of TIPP and TIPP-ψ was not simply due to the overexpression of δ-opioid receptors in GH₃ cells. The efficacy of TIPP and TIPP-ψ correlated with receptor density. This finding agrees with a previous study, which also found a correlation between receptor density and efficacy when comparing κ₁-opioid receptor binding and κ₁-receptor inhibition of adenylyl cyclase activity (Konkoy and Childers, 1993).

Importantly, not all the selective δ-antagonists investigated were capable of inhibiting adenylyl cyclase activity in GH₃DORT-8 cells. Under our conditions, naltriben and naloxone did not affect adenylyl cyclase activity and thus exhibited neutral antagonist activity. Previous studies have in fact reported naltriben and naloxone to be neutral antagonists (Szekeres and Traynor, 1997; Neilan et al., 1999). In contrast, naltrindole significantly inhibited adenylyl cyclase activity; however, this inhibition was not as efficacious as TIPP, TIPP-ψ, or the full δ-agonist DPDPE. Thus, naltrindole appeared to act as a partial agonist. Partial agonism by naltrindole has been reported in both in vitro and in vivo assays for δ-opioid receptor function (Stapelfeld et al., 1992; Szekeres and Traynor, 1997). The δ-antagonist ICI 174864 has been noted to exhibit inverse agonism (Szekeres and Traynor, 1997; Labarre et al., 2000), and this activity was observed in GH₃DORT-8 cells, where ICI 174864 significantly increased production of intracellular cAMP above control levels. Interestingly, analogs of the TIP(P) peptides containing the Dmt-Tic pharmacophore and HS378, an analog of naltrindole, have been shown to demonstrate partial to full inverse agonist activity (Labarre et al., 2000). Whether or not they are capable of inhibition of adenylyl cyclase activity remains to be determined.

Consistent with an agonist/antagonist interaction, ICI 174864 (1 μM) completely reversed the inhibition of adenylyl cyclase activity produced by 10 nM TIPP in GH₃DORT-8 cells. Surprisingly, this same concentration of ICI 174864 failed to completely reverse 0.5 nM DPDPE, although TIPP and DPDPE were shown to have similar affinities for the δ-opioid receptor. TIPP may bind differently to the receptor than DPDPE; thus, ICI 174864 can more efficiently compete against TIPP and reverse the effect. This seems possible based on recent studies investigating the δ-opioid receptor that have determined that each ligand/receptor interaction is unique and involves certain key residues that confer selectivity and affinity (Befort et al., 1996; Valiquette et al., 1996; Befort et al., 1999). Interestingly, TIPP potentiated, rather than reversed, the inhibition of adenylyl cyclase activity produced by DPDPE treatment. Consistent with results expected from an agonist/agonist interaction at the same receptor, our coadministration studies demonstrated an additive interaction between TIPP and DPDPE (Tallarida et al., 1989).

The possibility that the agonist activity of TIPP and TIPP-ψ was specific only for GH₃ cells was ruled out as both TIPP and TIPP-ψ significantly decreased the formation of intracellular cAMP in cells containing endogenous δ-opioid receptors (N1E115 and NG108-15) and in other cells containing transfected δ-opioid receptors (CHO-DOR and HEK-DOR).

In most of the cell lines examined in this study, the maximal decrease in forskolin-stimulated cAMP formation by TIPP was equivalent to that of DPDPE, suggesting TIPP acts as a full agonist. TIPP-ψ produced significant inhibition of adenylyl cyclase activity relative to control; however, this inhibition was significantly less than that produced by either TIPP or DPDPE, suggesting that TIPP-ψ acts as a partial agonist. In support of these hypotheses, comparison of receptor occupancy versus efficacy revealed that, in GH₃DORT-8 cells, both TIPP and DPDPE require <10% receptor occupancy to produce half-maximal inhibition of adenylyl cyclase activity (Table 2). In contrast, TIPP-ψ exhibits decreased maximal efficacy and requires ∼25% receptor occupancy for a similar level of inhibition of adenylyl cyclase activity. Full agonist activity of TIPP is also supported by the results of the coadministration studies. If TIPP had behaved as a partial agonist or an antagonist, the coadministration with DPDPE would have resulted in an antagonistic interaction. Instead, coadministration of TIPP with DPDPE resulted in an additive interaction. Thus, TIPP appears to act as a full agonist, comparable with DPDPE, whereas TIPP-ψ appears to act as a partial agonist.

Intracellular signaling of G-protein-coupled receptors is dependent upon the G-protein composition and stoichiometry, the downstream effectors present, and the presence of accessory proteins (Sato et al., 1995; Yang and Lanier, 1999). For example, clonidine can act as an agonist or partial agonist, depending on the final endpoint being examined (i.e., G-protein activation or effector regulation) and the tissue- or cell-specific architecture (Steer and Atlas, 1982;Surprenant et al., 1990). Therefore, it is possible that the agonist activity of TIPP and TIPP-ψ is only observed when the final endpoint examined is the intracellular effector adenylyl cyclase. This effect may be due to the selective activation of a single G-protein resulting in efficient coupling of the δ-opioid receptor to adenylyl cyclase. Previous studies showing δ-opioid receptors regulate adenylyl cyclase activity through G_iα2, and are more efficiently coupled to adenylyl cyclase than Ca²⁺ channels and support this hypothesis (McKenzie and Milligan, 1990; Prather et al., 2000). If TIPP and TIPP-ψ activated only a single G-protein subtype that was responsible for the regulation of adenylyl cyclase activity by δ-opioid receptors, then it is conceivable that this effect would not be detected by assays measuring GTPase activity or [³⁵S]GTPγS binding. Elucidating the exact mechanism(s) responsible for the agonist activity of TIPP and TIPP-ψ is the focus of current investigations in our laboratory.

The observation that TIPP and TIPP-ψ exhibit agonist activity has significant implications both in the research and clinical fields. It has been hypothesized that δ-opioid receptors play a role in the development of opioid tolerance and dependence (Abdelhamid et al., 1991). Studies have shown the concurrent administration of a δ-opioid receptor antagonist in mice chronically treated with morphine blocked the development of morphine tolerance and dependence without affecting morphine analgesia (Abdelhamid et al., 1991; Miyamoto et al., 1993). Similarly, equipotent doses of naltrindole, TIPP, and TIPP-ψ attenuated morphine tolerance and dependence in rats (Fundytus et al., 1995). This has led to the development of possible therapeutic opioid compounds with mixed μ-agonist/δ-antagonist properties displaying high analgesic efficacy, but significantly reduced tolerance and physical dependence. The discovery of agonist activity produced by TIPP and TIPP-ψ may lead to novel insights regarding the mechanism(s) of action underlying the development of opioid tolerance and dependence. These insights may in turn lead to a better understanding as to the exact role of δ-opioid receptors in the development of opioid tolerance and dependence.

In summary, the present study demonstrated the selective δ-opioid receptor antagonists TIPP and TIPP-ψ exhibited agonist activity by producing inhibition of adenylyl cyclase activity. This activity was selective for, and mediated by, activation of the δ-opioid receptor and could be reversed by selective δ-antagonists. In addition, the agonist effect of these compounds was concentration-dependent, independent of receptor density, and sensitive to PTX treatment. Furthermore, TIPP and TIPP-ψ inhibited adenylyl cyclase activity in cells that expressed both endogenous and transfected δ-opioid receptors. When coadministered, TIPP and DPDPE interacted in an additive manner to decrease formation of intracellular cAMP. The results of this study indicated that TIPP acted as a full agonist, comparable with DPDPE, whereas TIPP-ψ displayed partial agonist activity. These remarkable findings have broad implications for the future study of δ-opioid receptor signal transduction.