(2S,3R)β-Methyl-2′,6′-dimethyltyrosine-l-tetrahydroisoquinoline-3-carboxylic acid [(2S,3R)TMT-l-Tic-OH] Is a Potent, Selective δ-Opioid Receptor Antagonist in Mouse Brain

Materials and Methods

Chemicals.

(2S,3R)TMT-l-Tic-OH was synthesized at the University of Arizona, in the laboratory of Dr. V. J. Hruby (Liao et al., 1997). SNC80 and deltorphin II were obtained from Tocris Pharmaceuticals (Baldwin, MO); DAMGO was purchased from Sigma/RBI (Natick, MA). SNC80 was dissolved in 40% ethanol to obtain a 1 mM stock solution. Deltorphin II was initially prepared as 4 mM stock solution in 20% dimethyl sulfoxide. All other compounds were dissolved in distilled water (tail-flick test) or assay buffer ([³⁵S]GTPγS binding assay). Further dilutions were made by distilled water (tail-flick test) or assay buffer ([³⁵S]GTPγS binding assay). [³⁵S]GTPγS (1250 Ci/mmol) was purchased from PerkinElmer Life Sciences (Boston, MA).

Cell Membrane Preparation.

Male ICR mice supplied by Harlan Sprague-Dawley (Indianapolis, IN) were used in this study. The mice were sacrificed by cervical dislocation, whole brains were immediately removed, homogenized in 20 volumes of ice-cold TE buffer (10 mM Tris-HCl, 1.0 mM EDTA, pH 7.4) and centrifuged (40,000g, 15 min) to isolate the crude membrane fraction. Membranes were suspended in 20 volumes of ice-cold assay buffer (25 mM Tris-HCl, 150 mM NaCl, 2.5 mM MgCl₂, 1.0 mM EDTA, 50 μM GDP, 30 μM bestatin, 10 μM captopril, and 0.1 mM phenylmethylsulfonyl fluoride, pH 7.4) and incubated for 30 min at 30°C to remove endogenous agonists. The membranes were centrifuged again as described, and were resuspended in the assay buffer to an optical density of OD₂₈₀ = 0.8.

[³⁵S]GTPγS Binding Assay.

Mouse brain membranes were incubated with increasing concentrations of (2S,3R)TMT-l-Tic-OH in the presence of 0.1 nM [³⁵S]GTPγS (1250 Ci/mmol) in 1.0 ml of assay buffer (25 mM Tris, 150 mM NaCl, 50 μM GDP, 2.5 mM MgCl₂, 1 mM EDTA, 30 μM bestatin, 10 μM captopril, pH 7.4). To assess stimulation of [³⁵S]GTPγS binding by full δ- and μ-opioid receptor agonists in mouse brain membrane preparations, dose-response curves for SNC80 and DAMGO were determined as positive controls. Finally, dose-response curves were measured for SNC80 or DAMGO in the presence of (2S,3R)TMT-l-Tic-OH to determine K_e values for (2S,3R)TMT-l-Tic-OH at mouse brain δ- and μ-opioid receptors, respectively.

After a 90-min incubation at 30°C, the reaction was terminated by rapid filtration through Whatman GF/B glass fiber filters, followed by four rinses with 4 ml of ice-cold 25 mM Tris/120 mM NaCl, pH 7.4. The filters were soaked in 10 ml of EcoLite7 scintillation cocktail (ICN Biomedicals, Costa Mesa, CA) at 4°C for at least 6 h before bound radioactivity was measured by liquid scintillation spectrophotometry.

Warm (55°C) Water Tail-Withdrawal (Tail-Flick) Nociceptive Test in Mice.

Basal tail-withdrawal latencies were determined before administration of the opioid ligands. Mice with basal tail-withdrawal latencies longer than 5 s were excluded from further experiments. Each mouse received the appropriate dose of each compound in 5-μl solution volume, injected into the ventricles. Tail-withdrawal latencies were measured in 10-min intervals, 10 to 60 min after the injection, to determine an antinociceptive time course. Dose-response curves were determined at time points where the antinociceptive effect was maximal (20 min). Mice were assigned 100% antinociception when not responding within 15 s, and their tails were removed from the water to avoid tissue damage. Antinociception was calculated as maximal possible effect (MPE) using the formula: %MPE = (test latency − basal latency) × 100/(15-s basal latency).

Radiant Heat Paw-Withdrawal Hyperalgesic Test in Mice.

The method of Hargraeves et al. (1988) was used to determine whether (2S,3R)TMT-l-Tic-OH mediates hyperalgesia in mice. Mice (n = 8) were allowed to habituate to Plexiglas enclosures (2 × 3 inches) on a clear glass plate maintained at 30°C (Hargraeves box, purchased from Dr. T. Yaksh, University of California San Diego, CA). Baseline paw-withdrawal threshold values were determined by focusing a radiant heat source (high intensity projector lamp) onto the plantar surface of the hindpaw. Paw-withdrawal exposed a photocell that automatically turned off both the heat source and a timer. A cutoff of 30 s was used to prevent damage to the irradiated paw. (2S,3R)TMT-l-Tic-OH (30 nmol) was administed through i.c.v. route 1 h after the baseline paw-withdrawal test, and paw-withdrawal latencies were determined 15, 30, 45, and 60 min after the injection. A significant reduction in paw-withdrawal latency from the baseline value was interpreted as thermal hypersensitivity. Data were analyzed by one-way analysis of variance and significance was set at p ≤ 0.05.

Data Analysis.

[³⁵S]GTPγS binding data were analyzed as fixed (n_H = 1) slope sigmoidal dose-response curves using Prism Version 2 (San Diego, CA). The K_e value of (2S,3R)TMT-l-Tic-OH at the δ-opioid receptor was calculated using the Schild equation:K_e = [antagonist]/DR-1 (Arunlakshana and Schild, 1959), where DR indicates dose ratio between the EC₅₀ values of an agonist (SNC80) in the absence and presence, respectively, of an antagonist [(2S,3R)TMT-l-Tic-OH, 100 nM]. Since the dose-response of DAMGO was not significantly shifted in the presence of 3000 nM (2S,3R)TMT-l-Tic-OH, the exact K_e value of (2S,3R)TMT-l-Tic-OH at the μ-opioid receptor was not determined.

Results

We have reported previously (Hosohata et al., 1999) that (2S,3R)TMT-l-Tic-OH behaves as an inverse agonist in the functional [³⁵S]GTPγS binding assay in CHO cells overexpressing the human δ-opioid receptor. The present work was designed to determine the pharmacological properties of (2S,3R)TMT-l-Tic-OH in mouse brain using in vitro and in vivo functional assays.

The [³⁵S]GTPγS binding assay was used to investigate the effect of (2S,3R)TMT-l-Tic-OH on the nucleotide exchange rate of the G proteins in vitro in mouse brain membrane preparations. Opioid full agonists, selective for the δ- (SNC80) or μ- (DAMGO) opioid receptors (Quock et al., 1999), were used as controls to verify that functional receptor-G protein interaction is maintained in the mouse brain membrane preparations. As shown in Fig. 1, SNC80 and DAMGO stimulated [³⁵S]GTPγS binding in mouse brain membranes 150 ± 4.5 and 152 ± 5.7% over the basal level, respectively. The EC₅₀ values of SNC80 and DAMGO were 246 ± 20 and 1460 ± 630 nM, respectively. In contrast, (2S,3R)TMT-l-Tic-OH up to 30 μM concentration did not influence basal [³⁵S]GTPγS binding, indicating that (2S,3R)TMT-l-Tic-OH is a neutral antagonist in mouse brain membrane preparations (Fig. 1).

Additionally, (2S,3R)TMT-l-Tic-OH dose dependently shifted to the right of the dose-response curve of SNC80 in the [³⁵S]GTPγS binding assay (Fig.2A). The potency of (2S,3R)TMT-l-Tic-OH at the δ-opioid receptor, as determined using the Schild equation, wasK_e = 3.61 ± 0.71 nM. In contrast, (2S,3R)TMT-l-Tic-OH had no significant effect on the dose-response curve of the μ-opioid agonist, DAMGO, in the [³⁵S]GTPγS binding assay in mouse brain membranes (Fig. 2B). Therefore, theK_e value of (2S,3R)TMT-l-Tic-OH at the δ-opioid receptor can be estimated to be at least 800-fold lower than its K_e value at the μ-opioid receptor (>3000) in mouse brain.

The in vivo properties of (2S,3R)TMT-l-Tic-OH were investigated using the warm (55°C) water tail-withdrawal (tail-flick) nociceptive and the radiant heat paw-withdrawal hyperalgesic tests in mice. As seen in Fig. 3, i.c.v. injection of δ- (deltorphin II, Fig. 3A) or μ- (DAMGO, Fig. 3B) opioid receptor agonists increased tail-flick latencies in mice. Antinociception was maximal 20 min after i.c.v. administration of either agonist, in doses ≥10 nmol/mouse for deltorphin II and ≥0.02 nmol/mouse for DAMGO. In contrast, as seen in Fig.4, (2S,3R)TMT-l-Tic-OH individually had no measurable effect on tail-withdrawal latencies. The lack of effect of (2S,3R)TMT-l-Tic-OH persisted throughout the examined time interval (60 min) after i.c.v. administration of the compound (Fig. 4) for doses up to 30 nmol/mouse.

Putative hyperalgesic properties of (2S,3R)TMT-l-Tic-OH were also tested in the more sensitive radiant heat paw-withdrawal test in mice (Hargraeves et al., 1988). The mean paw-withdrawal threshold in vehicle-injected control mice was 10.4 ± 0.6 s. Intracerebroventricular injection of (2S,3R)TMT-l-Tic-OH (30 nmol/mouse) had no measurable effect (one-way analysis of variance,p = 0.42, nonsignificant) on paw-withdrawal latencies in this highly sensitive hyperalgesic test (Fig.5). An established cannabinoid CB₁ receptor inverse agonist (SR141,716) was used as a positive control (Richardson et al., 1997) in the assay. We found that SR141,716 (1 nmol/mouse) reduced paw-withdrawal latencies by approximately 50% 5 min after i.c.v. administration (data not shown), indicating that the method is sensitive enough to detect hyperalgesia by an established CB₁ receptor inverse agonist (Landsman et al., 1997).

Finally, the selectivity of (2S,3R)TMT-l-Tic-OH (30 nmol/mouse) to antagonize antinociception mediated by δ- or μ-opioid receptor agonists was also investigated. To establish agonist doses necessary and sufficient to achieve maximal possible antinociceptive effect (MPE) in the tail-flick test 20 min after i.c.v. adminstration, dose-response curves were determined for the δ-selective agonist deltorphin II or the μ-selective agonist DAMGO. Both deltorphin II and DAMGO produced maximal antinociception 20 min after i.c.v. injection, with A₅₀values of 8.55 and 0.074 nmol/mice, respectively. Therefore, in subsequent experiments, mice received i.c.v. injection of a high dose of (2S,3R)TMT-l-Tic-OH (30 nmol/mouse) 10 min prior to i.c.v. administration of the previously determined optimal doses of either deltorphin II (20 nmol/mouse) or DAMGO (0.2 nmol/mouse). Tail-flick latencies were determined 20 min after the injection of the agonists. As shown in Fig.6, pretreatment of mice with 30 nmol (2S,3R)TMT-l-Tic-OH reduced deltorphin II-mediated antinociception (40 ± 13% MPE,p < 0.05) but had no effect on DAMGO-mediated antinociception.

Discussion

Selective δ-opioid receptor antagonists can be used as pharmacological tools to identify unique physiological functions of the δ-opioid receptors and may subsequently serve as therapeutic agents to regulate those functions in vivo. Selective δ-opioid antagonists were shown to modulate the development of tolerance and dependence to opiate agonists (Abdelhamid et al., 1991; Fundytus et al., 1995) and to diminish the reinforcing effects of cocaine (Menkens et al., 1992) and ethyl alcohol (Froehlich et al., 1998). δ-Opioid antagonists may also have therapeutic value as immunosuppressants in organ transplants (House et al., 1995) and in chronic inflammatory diseases, such as rheumatoid arthritis (Spetea et al., 2001).

Structure-activity studies on morphinane- and enkephalin-based opioids led to the development of nonpeptidic and peptidic opioid antagonists with improved selectivity for the δ-opioid receptor. Thus, introduction of steric constraints into the prototype δ-opioid peptide antagonist, N,N-diallyl-Leu-enkephalin (Shaw et al., 1982), yielded metabolically stable pseudopeptide antagonists, such as ICI 174,864 (Cotton et al., 1984), TIPP (Schiller et al., 1992), TIPPΨ (Schiller et al., 1993), and Tyr-Tic (Salvadori et al., 1995), with progressively increasing δ-selectivity. In the second generation of pseudopeptide opioid antagonists, additional steric constraints were introduced by the use of methylated Tyr analogs, leading to further improved δ-selectivity, such as DMT-l-Tic (Salvadori et al., 1995, 1997) and TMT-l-Tic (Liao et al., 1997).

We have previously isolated the stereoisomers of the highly constrained opioid dipeptide TMT-l-Tic-OH to study the effects of steric and topographic constraints on the affinity, selectivity, and functional properties of the opioid pharmacophore. The most selective stereoisomer, (2S,3R)TMT-l-Tic-OH, competed with high affinity (IC₅₀ 9.3 ± 0.53 nM) for [³H]p-Cl-DPDPE, but not for [³H]DAMGO (IC₅₀35,000 ± 18,000) binding sites in rat brain membrane preparations (Liao et al., 1997). Similar affinity and selectivity values were obtained for (2S,3R)TMT-l-Tic-OH in recombinant CHO cell lines, expressing human δ- or μ-opioid receptors (K_i = 5.2 ± 0.8 and 33,600 ± 4,340 nM, respectively; unpublished data).

Subsequently, we reported that (2S,3R)TMT-l-Tic behaves as a potent full inverse agonist in a recombinant Chinese hamster ovary cell line (hDOR/CHO), stably expressing a high number of human δ-opioid receptors per cell (1800 ± 150 fmol/million cells). (2S,3R)TMT-l-Tic-OH inhibited basal [³⁵S]GTPγS binding in hDOR/CHO cells with an intrinsic activity similar to that of the prototype δ-opioid receptor inverse agonist, ICI 174,864 (E_max = 42 versus 47.8% of basal, respectively), but had 78-fold higher potency (Hosohata et al., 1999).

According to the two-state ternary complex model of receptor activation (Samama et al., 1993), G protein-coupled receptors exist in equilibrium between inactive and active receptor conformations. The active receptor conformation is able to activate G proteins, leading to signal transduction. The propensity of a receptor to convert from inactive to active conformation in the absence of ligand (constitutive activity) is determined by the structure of the receptor. Some receptors have been shown to exhibit measurable constitutive activity in physiological tissue preparations (reviewed by de Ligt et al., 2000), whereas for others spontaneous isomerization to the active state is more strongly restricted.

The ability of different ligands to perturb the equilibrium between inactive and active receptor conformations is described by a ligand-specific equilibrium constant β, the microscopic equivalent of efficacy (Onaran and Costa, 1997). Ligands with β values >1 are agonists that shift the equilibrium into the direction of the active ternary complex. Inverse agonists (β < 1) have higher affinity to the inactive receptor conformation and will shift the equilibrium in the opposite direction, whereas neutral agonists (β = 1) have no effect on the equilibrium.

Experimental manipulations, such as overexpression of the receptor or the appropriate G proteins, or the presence of low sodium or GDP concentrations in the assay buffer, were shown to increase constitutive receptor activity (Strange, 2002). In addition, a number of point mutations have been identified that increased constitutive GPCR activity. Some of these mutants have been implicated in the pathophysiology of diseases (reviewed in de Ligt et al., 2000). Recently, a number of drugs that were previously considered neutral antagonists were shown to exhibit inverse agonist properties in recombinant cells expressing very high receptor densities.

The relevance of inverse agonism in tissues expressing lower physiological receptor concentrations, however, is still not fully understood (de Ligt et al., 2000). Therefore, it is of interest to examine the pharmacological properties of (2S,3R)TMT-l-Tic-OH in the central nervous system to determine its mechanism of action in vivo. In the present study, we examined the pharmacological properties of (2S,3R)TMT-l-Tic-OH in mouse brain. Interestingly, (2S,3R)TMT-l-Tic-OH behaved as a neutral δ-opioid receptor antagonist in mouse brain, exhibiting high potency in vitro in the [³⁵S]GTPγS binding assay and high selectivity in both in vitro and in vivo (warm water tail-flick) pharmacological assays.

Thus, (2S,3R)TMT-l-Tic-OH did not influence basal [³⁵S]GTPγS binding in mouse brain membranes. (2S,3R)TMT-l-Tic-OH also had no significant effect on tail withdrawal latencies when administered alone. Similar to (2S,3R)TMT-l-Tic-OH, the prototype δ-opioid receptor inverse agonist ICI 174,864, also had no effect on basal [³⁵S]GTPγS binding in rat cerebral cortex and striatum cell membranes (Rouleau et al., 2002), indicating that the expression level of the δ-opioid receptor in whole brain homogenates does not reach the threshold level to produce measurable constitutive activity in this assay. Interestingly however,Georgoussi and Zioudrou (1993), using ³²P as radiolabel, were able to show a 10% inhibition of high affinity GTPase activity by 1 μM ICI 174,864 in rat brain membranes.

Physiological tissue preparations usually consist of mixed populations of cells, each expressing different receptor concentrations. Thus, although (2S,3R)TMT-l-Tic-OH did not influence basal G protein activity in whole brain homogenates and showed no nociceptive response in the tail-flick assay, it is still conceivable that a more sensitive assay would detect inverse agonism in responses mediated in critical brain regions that express high δ-opioid receptor density. Nociceptive tests based on changes in paw-withdrawal latency in response to thermal nociceptive stimuli have proven to be more sensitive for the detection of hyperalgesic responses than the warm water tail-flick assay (Richardson et al., 1997). Therefore, we also used the radiant heat paw-withdrawal latency assay to test whether (2S,3R)TMT-l-Tic-OH exhibited any inverse agonist (hyperalgesic) properties in vivo in mice. Intracerebroventricular injection of (2S,3R)TMT-l-Tic-OH (30 nmol/mouse), however, had no measurable effect on paw-withdrawal latencies in the highly sensitive radiant heat paw-withdrawal hyperalgesic test. The CB₁ cannabinoid receptor inverse agonist, SR 141,716, on the other hand was able to produce hyperalgesia in this test. The difference is likely due to the 10 times higher density of CB₁ receptors in brain.

(2S,3R)TMT-l-Tic-OH on the other hand dose dependently shifted the dose-response curve of a δ-selective opioid receptor agonists in the [³⁵S]GTPγS binding assay with aK_e value of 3.6 ± 0.7 nM. Conversely, the dose-response curve of the μ-selective opioid agonist, DAMGO was not altered in the presence of (2S,3R)TMT-l-Tic-OH. Moreover, i.c.v. injection of (2S,3R)TMT-l-Tic-OH 10 min before administration of opioid receptor agonists antagonized deltorphin II-mediated antinociception in the warm water tail-withdrawal test in mice but had no effect on DAMGO-mediated antinocicepton even at a high (30 nmol) dose.

These results indicate that (2S,3R)TMT-l-Tic-OH is a potent highly selective neutral δ-opioid antagonist in mouse brain membrane preparations in vitro and a selective neutral δ-opioid antagonist in mouse brain in vivo. (2S,3R)TMT-l-Tic therefore, may serve as a useful pharmacological and therapeutic agent, to characterize and modulate δ-opioid receptor-regulated physiological functions.