Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The development of tolerance remains a significant problem with the use of opioid analgesics for the treatment of chronic pain. Morphine and other clinically available opiates act primarily at the µ-opioid receptor, but they also have affinity for - and -opioid receptors. Blockade of -receptors with -selective opioid antagonists greatly reduced the development of morphine tolerance and dependence in mice (Abdelhamid et al., 1991; Miyamoto et al., 1993; Fundytus et al., 1995). Reduction of -receptors with antisense oligodeoxynucleotides reduced morphine tolerance (Kest et al., 1996), and there was a loss of morphine tolerance in -receptor knockout mice (Zhu et al., 1999). These studies point to a role for -receptors in the development of opioid tolerance and led one of us (P.W.S.) to propose the development of compounds that have mixed µ-agonist and -antagonist activities. The first of such mixed µ-agonist/-antagonist to be tested in vivo was (H-Dmt-Tic[CH₂NH]Phe-Phe-NH₂; DIPP-NH₂[]) (Schiller et al., 1999). DIPP-NH₂[] was found to be three times more potent than morphine after i.c.v. administration. Continuous infusion of DIPP-NH₂[] produced less acute tolerance than morphine, but higher doses produced a similar degree of tolerance as morphine. The results of this latter study led us to question whether activation of the -receptor is necessary for the development of tolerance or whether the -receptor is only involved in the state of tolerance.

To address this question, we used an extraordinarily selective µ-agonist, [Dmt¹]DALDA (H-2',6'dimethyltyrosine-D-Arg-Phe-Lys-NH₂). [Dmt¹]DALDA is a dermorphin analog with extraordinary selectivity for the µ-receptor (K_i/K_i^µ > 14,000 and K_i/K_i^µ > 100) (Schiller et al., 2000). [Dmt¹]DALDA has high affinity for the µ-receptor (K_i = 143 pM) and was found to be extremely potent after intrathecal (i.t.) administration in rats (ED₅₀ = 1 pmol) (Shimoyama et al., 2001). [Dmt¹]DALDA is 3000 times more potent compared with morphine at the spinal cord. At equipotent doses, the duration of action of intrathecal [Dmt¹]DALDA was 12 h compared with 3 h for morphine. This long duration of action is believed to be due to its slow clearance from spinal fluid because of its 3+ net charge at physiological pH. Surprisingly, this very polar peptide analog was found to be even more potent than morphine after subcutaneous (s.c.) administration (Neilan et al., 2001).

Most interestingly, there seems to be no cross-tolerance between [Dmt¹]DALDA and morphine. In mice treated with escalating doses of s.c. morphine for 4 days, there was no significant change in the ED₅₀ for s.c. [Dmt¹]DALDA (Neilan et al., 2001). This lack of cross-tolerance between [Dmt¹]DALDA and morphine suggests either different receptor mechanisms for the two drugs or that there is a very low propensity for tolerance to [Dmt¹]DALDA because of its lack of action at the -receptor. Based on antisense studies, Neilan and colleagues (2001) proposed that unlike morphine and other µ-agonists, the action of [Dmt¹]DALDA does not seem to be mediated via MOR-1. However, the development of tolerance to [Dmt¹]DALDA itself has never been examined.

In this study, we examined the development of tolerance to [Dmt¹]DALDA after repeated subcutaneous administration. To our surprise, we found rapid onset of tolerance after as few as five doses of [Dmt¹]DALDA, and tolerance was profound in the spinal cord. We also found evidence that animals tolerant to [Dmt¹]DALDA were also tolerant to morphine. Finally, although the -opioid receptor plays no role in the analgesic action of [Dmt¹]DALDA in the spinal cord, administration of a -antagonist partially restored potency to intrathecal [Dmt¹]DALDA in the tolerant animal.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals

Male CD-1 mice (25-30 g) were purchased from Charles River Laboratories (Wilmington, MA). Animals were housed in a temperature-controlled room maintained on a 12-h-light/dark cycle. Food and water were available ad libitum until the time of the experiment. All experiments were conducted in accordance with the ethical guidelines of the International Association for the Study of Pain and approved by the Institution for the Care and Use of Animals at Weill Medical College of Cornell University. Male Sprague-Dawley rats (300-350g) were used in some experiments, and these were conducted in accordance with guidelines approved by the Institutional Animal Use Committee of Chiba University School of Medicine.

Drugs and Chemicals

[Dmt¹]DALDA and H-Tyr-Tic[CH₂NH]Phe-Phe-OH (TIPP[]) were synthesized according to methods described previously (Schiller et al., 1993, 2000). Morphine sulfate and [D-Ala²]-deltorphin I (DELT) were supplied by the National Institute on Drug Abuse (Rockville, MD). For the preparation of [Dmt¹]DALDA in tritiated form, a precursor peptide containing 2',6'-dimethyl-3',5'-diiodotyrosine [Tyr(2',6'-Me₂-3',5'-I₂)] needed to be synthesized. Fmoc-Dmt-OH was iodinated by treatment with I₂ in the usual manner to yield Fmoc-Tyr(2',6'-Me₂-3',5'-I₂)-OH. This protected amino acid was then used in the solid phase synthesis of H-Tyr(2',6'-Me₂-3',5'-I₂)-D-Arg-Phe-Lys-NH₂ according to a protocol published elsewhere (Schiller et al., 2000). The peptide was purified by preparative reversed-phase chromatography, and its structure was confirmed by fast atom bombardment mass spectrometry. Catalytic tritiation of this precursor peptide was performed at the Institute of Isotopes (Budapest, Hungary), resulting in a product with a specific radioactivity of 47.18 Ci/mmol. All other drugs and chemicals were obtained from Sigma-Aldrich (St. Louis, MO).

Drug Administration

Mice Studies. Drugs were injected intrathecally (i.t.) according to the method described by Hylden and Wilcox (1980). The needle (30 gauge) was inserted from the side of the L5 or L6 spinous process, and the injection volume was 4 µl/mouse. For intracerebroventricular (i.c.v.) injections, mice were lightly anesthetized with isoflurane, and an incision was made over the scalp to expose the bregma. The injection (4 µl) was delivered 2 mm lateral and caudal to the bregma to a depth of 3 mm (Haley and McCormick, 1957).

Rat Studies. Intrathecal catheterization was performed at least 2 days before the experiments as previously described (Shimoyama et al., 1996, 2001). In brief, under halothane anesthesia, a PE-10 tube was inserted through a small hole made in the atlanto-occipital membrane and threaded 8.5 cm down the intrathecal space to the lumbosacral level of the spinal cord. Drugs were delivered via the catheter in a volume of 5 µl, followed by 10 µl of saline to flush the catheter.

Antinociceptive Assay

The radiant heat tail-flick assay was used for antinociceptive tests in mice and rats. The light intensity was adjusted such that the baseline latencies ranged between 2.5 and 3.5 s. Analgesia was defined as a latency response of greater than two times the baseline latency for an individual animal. To avoid tissue damage, a cut-off of 10 s was used in mice, and a cut-off of 7 s was used in rats. In all antinociceptive tests in mice, groups of 8 to 10 mice were used for each dose, and each mouse was only used once. The percentage of analgesic responders was calculated, and the quantal dose-response curves analyzed using probit analysis (PharmTools Pro; McCary Group, Inc., Elkins Park, PA). Data are presented as ED₅₀ with 95% confidence intervals. Statistical comparisons of dose-response curves were performed by analysis of variance with F-statistic (PharmTools Pro).

Study 1. Analgesic Potency of [Dmt¹]DALDA and Morphine after s.c., i.t., and i.c.v. Administration. [Dmt¹]DALDA and morphine were administered s.c., i.t., or i.c.v. to naive mice. Potency was determined at time of peak effect as determined from preliminary time course data. Time course of analgesia was then determined using 5 times the ED₅₀ dose of [Dmt¹]DALDA and morphine.

Study 2. Role of -Receptors in Spinal Action of [Dmt¹]DALDA. To demonstrate lack of [Dmt¹]DALDA activity at -receptors, mice were given naltriben (NTB; 3 mg/kg, s.c.) 30 min before i.t. [Dmt¹]DALDA or DELT.

Study 3. Tolerance and Cross-Tolerance to [Dmt¹]DALDA and Morphine. In the first experiment, mice were treated with s.c. [Dmt¹]DALDA (5 times ED₅₀, every day) for a total of 7 days. Analgesic potency of s.c. [Dmt¹]DALDA and s.c. morphine were determined on day 8 in both groups of animals. In the second experiment, mice were treated with [Dmt¹]DALDA twice a day (10 times ED₅₀, every 12 h) for a total of 2.5 days. Analgesic potency of [Dmt¹]DALDA and morphine were then assessed after s.c. administration on day 4. In the third experiment, mice were treated twice daily with [Dmt¹]DALDA (10 times ED₅₀, every 12 h) for a total of 2.5 days, and analgesic potency of [Dmt¹]DALDA was assessed after i.c.v. and i.t. administration. In addition, brains and spinal cords were removed from these animals after repeated dosing for radioligand binding assays.

Study 4. Tolerance to [Dmt¹]DALDA after Repeated i.t. Injections in Rats. The finding of profound spinal tolerance to [Dmt¹]DALDA in mice after repeated s.c. [Dmt¹]DALDA administration was surprising and led us to investigate whether tolerance would develop to repeated i.t. [Dmt¹]DALDA administration. Repeated i.t. injections were carried out in rats because it was technically difficult to give i.t. injections to the same mouse twice a day for several days. Cumulative dose-response studies were, therefore, performed with groups of 10 rats. After measuring the baseline latencies, increasing doses of [Dmt¹]DALDA were administered via the intrathecal catheter until each animal in the group became an analgesic responder (Shimoyama et al., 2001). Rats were then treated repeatedly with i.t. [Dmt¹]DALDA. One group received 10 pmol (10 times ED₅₀) daily for 3 days, a second group received 10 pmol twice a day for 3 days, and the cumulative dose-response curve repeated on day 4. A third group of rats received 10 pmol of [Dmt¹]DALDA (i.t.) daily for 7 days, and the dose-response curve was repeated on day 8.

Study 5. Tolerance to i.t. [Dmt¹]DALDA after Repeated s.c. Injection together with a -Antagonist. Mice were treated with [Dmt¹]DALDA (10 times ED₅₀) and NTB (3 or 5 mg/kg) concurrently, given s.c. every 12 h, for a total of 2.5 days. Analgesic potency of [Dmt¹]DALDA was then assessed after i.t. administration on day 4.

Study 6. Effect of -Antagonists on Analgesic Potency of i.t. [Dmt¹]DALDA in Tolerant Mice. Mice were pretreated with s.c. [Dmt¹]DALDA (10 times ED₅₀, every 12 h for 2.5 days). Analgesic potency of i.t. [Dmt¹]DALDA was determined on day 4 either alone, after pretreatment with NTB (3 mg/kg, s.c., given 30 min before [Dmt¹]DALDA), or together with TIPP[] (0.2 nmol or 2 nmol, i.t.).

Radioligand Binding Assay

Mice were decapitated, and brain and spinal cord were removed and put in 30 volumes of ice-cold 50 mM Tris buffer (pH 7.4) immediately. Tissues were homogenized for 10 s and centrifuged at 40,000g for 20 min at 4°C. After this process was repeated a second time, the pellet was re-suspended in Tris buffer (pH 7.4) at a final concentration of 2 mg/ml. Protein concentrations were determined by the Bradford procedure (Bio-Rad, Hercules, CA). Aliquots of membrane homogenates (400 µg of protein) were incubated with [³H][Dmt¹]DALDA (10-800 pM) for 60 min at 25°C. Nonspecific binding was assessed by inclusion of 1 µM unlabeled [Dmt¹]DALDA. Free radioligand was separated from bound radioligand by rapid filtration through GF/B filters (Brandel, Gaithersberg, MD) with a cell harvester (Brandel). Filters were washed 3 times with 10 ml of Tris buffer. Radioactivity was determined by liquid scintillation counting. Binding affinities (K_d) and receptor number (B_max) were determined using nonlinear regression (GraphPad, San Diego, CA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Analgesic Potency of [Dmt¹]DALDA and Morphine after s.c., i.t., and i.c.v. Administration. [Dmt¹]DALDA was more potent than morphine after all three routes of administration (Fig. 1). The ED₅₀ values for [Dmt¹]DALDA and morphine are summarized in Table 1. [Dmt¹]DALDA was most potent after i.t. administration, being 833 times more potent compared with morphine. Even when administered s.c., [Dmt¹]DALDA was 36-fold more potent than morphine. Whereas morphine was equipotent after i.c.v. and i.t. administration, [Dmt¹]DALDA was 3.8 times more potent after i.t. administration.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Dose-dependent antinociceptive effects of s.c., i.c.v., and i.t. [Dmt1]DALDA () and morphine () in the mouse tail-flick assay (n = 8-10 in each group). Dose-response curves were determined at time of peak effect (30 min after s.c., i.c.v., or i.t. administration). ED50 and 95% confidence intervals are given in Table 1.

Time Course of [Dmt¹]DALDA after s.c. and i.t. Administration. [Dmt¹]DALDA produced prolonged analgesia after both s.c. and i.t. administration (Fig. 2). A significant increase in tail-flick latency was observed for 12 h with a s.c. dose that is equivalent to 5 times ED₅₀. In contrast, the antinociceptive response to an equipotent dose of morphine was only 3 h.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Time course of antinociceptive effects of subcutaneous [Dmt1]DALDA () and morphine () (left panel) and intrathecal [Dmt1]DALDA () (right panel). A dose of 5 times ED50 value was given. Tail-flick latencies were measured before and at various times after drug administration (n = 8-10 in each group).

Effect of -Antagonists on Spinal Action of [Dmt¹]DALDA. NTB significantly shifted the dose-response curve for i.t. DELT to the right, with ED₅₀ increasing from 7.4 (4.0-15.3) to 20.8 (12.1-40.9) nmol/mouse. In contrast, NTB had no effect on the ED₅₀ of i.t. [Dmt¹]DALDA [1.4 (0.8-2.7) versus 2.5 (0.8-5.3) pmol/mouse].

Tolerance and Cross-Tolerance after s.c. [Dmt¹]DALDA Administration. Daily s.c. injection of [Dmt¹]DALDA (5 times ED₅₀) for 7 days resulted in a 3.6-fold shift in the dose-response curve for [Dmt¹]DALDA when tested on day 8 (Fig. 3). The ED₅₀ values are summarized in Table 2. Twice daily dosing at 10 times ED₅₀ for just 2.5 days resulted in a 11.7-fold shift in the dose-response curve on day 4. Thus, dose seemed to be a more important factor in determining the magnitude than dosing frequency. Twice daily dosing of [Dmt¹]DALDA (10 times ED₅₀) for 2.5 days produced a much greater shift in the i.t. (44.1-fold) compared with s.c. (11.7-fold) or i.c.v. (3.3-fold) [Dmt¹]DALDA dose-response curves when tested on day 4 (Table 3). Mice pretreated with [Dmt¹]DALDA were also tolerant to morphine (Fig. 3). The two [Dmt¹]DALDA pretreatment paradigms produced a corresponding 3.4- and 15.1-fold shift in the morphine dose-response curve (Table 2). The relative potency of [Dmt¹]DALDA to morphine remained the same in the tolerant mice. It should be pointed out that in all of these experiments, dose-response curves obtained after repeated [Dmt¹]DALDA administration were compared with dose-response curves obtained in naïve animals rather than animal that had received repeated saline treatment.

View larger version (9K): [in this window] [in a new window]   Fig. 3.   Dose-response curves for [Dmt1]DALDA (left panel) and morphine (right panel) in naïve mice (filled symbols) and mice treated repeatedly with s.c. [Dmt1]DALDA (5 times ED50, daily for 7 days) (open symbols) or s.c. [Dmt1]DALDA (10 times ED50, twice daily for 3 days) (gray symbols). Dose-response curves were determined 30 min after s.c. [Dmt1]DALDA or morphine administration. ED50 and 95% confidence intervals are given in Table 2.

Tolerance to [Dmt¹]DALDA after Repeated i.t. Administration in Rats. The profound spinal tolerance to [Dmt¹]DALDA observed in the mouse after s.c. administration was confirmed by repeated i.t. injections in rats. Treatment of rats with i.t. [Dmt¹]DALDA (10 times ED₅₀) once a day for 3 days produced a 4.2-fold shift in the i.t. dose-response curve on day 4 (Fig. 4). Administration of the same dose of [Dmt¹]DALDA twice a day resulted in a 43-fold shift in the dose-response curve. Tail-flick latency decreased progressively over a 7-day treatment with i.t. [Dmt¹]DALDA (10 times ED₅₀) so that the antinociceptive response was abolished by day 5. Baseline latency on day 1 was 2.9 ± 0.16 s, and a dose response was observed with i.t. [Dmt¹]DALDA on day 1. On day 8, baseline latency was 2.16 ± 0.28 s, and it was no longer possible to obtain a dose-response curve for i.t. [Dmt¹]DALDA even at 1000 times the original ED₅₀.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Left panel, dose-response curves for [Dmt1]DALDA in naive rats (filled symbols) and rats treated repeatedly with i.t. [Dmt1]DALDA (10 times ED50, daily for 3 days) (open symbols) or i.t. [Dmt1]DALDA (10 times ED50, twice daily for 3 days) (gray symbols). Right panel, dose-response curves for i.t. [Dmt1]DALDA on day 1 (filled symbols) and day 8 (gray symbols) during daily i.t. treatment with [Dmt1]DALDA (10 times ED50).

Tolerance to i.t. [Dmt¹]DALDA after Repeated s.c. Injection together with a -Antagonist. Concurrent administration of 3 mg/kg NTB with [Dmt¹]DALDA twice a day for 2.5 days significantly reduced the tolerance observed when i.t. [Dmt¹]DALDA was tested on day 4 (Table 4). There was no significant difference between the two doses of NTB. NTB had no effect on the ED₅₀ of i.t. [Dmt¹]DALDA in naive animals (see above).

Effect of -Antagonists on i.t. [Dmt¹]DALDA Analgesic Potency in Tolerant Mice. The addition of -antagonists significantly enhanced i.t. [Dmt¹]DALDA potency in mice already made tolerant by 2.5 days of repeated s.c. [Dmt¹]DALDA. Administration of NTB (3 mg/kg, s.c.) 30 min before i.t. [Dmt¹]DALDA significantly shifted the [Dmt¹]DALDA dose-response curve to the left in tolerant mice (Table 5). Concurrent administration of TIPP[] (0.2 nmol, i.t.) partially restored the ED₅₀ of i.t. [Dmt¹]DALDA from 52.9 to 22.6 pmol/mouse. A higher dose of TIPP[] (2.0 nmol) resulted in a further decrease in ED₅₀ (11.76 pmol/mouse). TIPP[] had no effect on [Dmt¹]DALDA potency in naive mice (data not shown).

View this table: [in this window] [in a new window]   TABLE 5 Effects of  antagonists on i.t. [Dmt1]DALDA antinociceptive potencies in tolerant mice All mice were treated with s.c. [Dmt1]DALDA 1.6 µmol/kg (10 times ED50) twice a day for 2.5 days. Antinociceptive tests were performed on day 4. Naltriben (NTB) was administered s.c. 30 min before i.t. [Dmt1]DALDA. TIPP[] was administered together with [Dmt1]DALDA i.t.

[³H][Dmt¹]DALDA Binding in Naive and Tolerant Mouse Brain and Spinal Cord. Figure 5 illustrates the saturation binding curves for [³H][Dmt¹]DALDA in mouse brain membrane homogenates. High-affinity specific binding was observed in both brain and spinal cord membranes, with K_d ~125 to 150 pM. The number of binding sites was higher in brain compared with spinal cord. The binding of [³H][Dmt¹]DALDA to brain membranes was completely displaced by morphine (K_i = 5.64 ± 0.24 nM; n = 4) and DAMGO (K_i = 2.80 ± 0.14 nM; n = 4). Table 6 summarizes the K_d and B_max values in naive and tolerant brain and spinal cord membranes. Treatment with s.c. [Dmt¹]DALDA (10 times ED₅₀) twice a day for 2.5 days resulted in 30 to 35% reduction in B_max in both brain and spinal cord with no change in binding affinity. [Dmt¹]DALDA binding was not affected by the presence of TIPP[] either in naive or tolerant tissues (data not shown).

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Left panel, saturation binding curve for [3H][Dmt1]DALDA using mouse brain membranes. Right panel, competitive displacement of [3H][Dmt1]DALDA binding with morphine and DAMGO. Membrane homogenates were incubated with [3H][Dmt1]DALDA (10-800 pM) for 60 min at 25°C. Nonspecific binding was assessed in the presence of 1 µM unlabeled [Dmt1]DALDA. The Kd and Bmax values are summarized in Table 6.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Recent findings suggesting that the -receptor plays an important role in the development of morphine tolerance led us to hypothesize that tolerance may be minimized with a highly selective µ-agonist. [Dmt¹]DALDA has been shown to be 14,000 times more selective for the µ-receptor compared with -receptor (Schiller et al., 2000). The results of the present study, however, actually showed rapid onset of tolerance with [Dmt¹]DALDA, and tolerance was profound in the spinal cord.

The high potency of s.c. [Dmt¹]DALDA was unexpected because of its very polar nature (due to its 3+ net charge at physiological pH) and presumed difficulty in crossing the blood-brain barrier. Another unexpected finding was its long duration of action after s.c. administration. When [Dmt¹]DALDA was given s.c. at 5 times the ED₅₀, a significant increase in tail-flick latency was observed for 12 h. Administration of this dose once a day for 7 days resulted in a 3.6-fold rightward shift in the dose-response curve. Twice daily dosing with a higher dose (10 times ED₅₀) of [Dmt¹]DALDA resulted in a 11.7-fold shift in the dose-response curve after as few as five doses. In contrast, repeated injections of morphine in a similar dose and dosing interval produced only a 2-fold increase in the ED₅₀ (data not shown).

As expected, [Dmt¹]DALDA was much more potent after i.c.v. or i.t. administration. Unlike morphine, which is equipotent after i.c.v. and i.t. administration, [Dmt¹]DALDA was ~4 times more potent after i.t. compared with i.c.v. administration. In this study, i.t. [Dmt¹]DALDA was ~800-fold more potent than i.t. morphine. Previous studies reported an i.t. potency ratio of 3000 in rats (Shimoyama et al., 2001) and 5000 in mice (Neilan et al., 2001). The higher potency of [Dmt¹]DALDA at the spinal cord may, in part, be explained by its ability to inhibit norepinephrine uptake, resulting in a synergistic relationship between µ-receptors and ₂-adrenergic receptors in the spinal cord (Shimoyama et al., 2001). Along with the higher potency of [Dmt¹]DALDA in the spinal cord, the magnitude of spinal tolerance was far greater than supraspinal tolerance after s.c. administration of [Dmt¹]DALDA. The ED₅₀ for i.t. [Dmt¹]DALDA was increased 44-fold, whereas the ED₅₀ for i.c.v. [Dmt¹]DALDA was only increased by 3-fold. Tolerance to s.c. [Dmt¹]DALDA (11.7-fold) was intermediate between spinal and supraspinal tolerance, suggesting that the spinal cord, rather than the brain, is likely to be the primary site of action of [Dmt¹]DALDA after systemic administration. The profound spinal tolerance observed in mice after systemic [Dmt¹]DALDA treatment was confirmed by studies using repeated i.t. administration in rats. Intrathecal injection of [Dmt¹]DALDA once a day for 7 days resulted in a complete loss of antinociceptive response to i.t. [Dmt¹]DALDA.

The reason behind the discrepancy between spinal and supraspinal tolerance is not clear. One possibility is that [Dmt¹]DALDA does not distribute appreciably into the brain after systemic administration. There are anatomical differences between the blood-brain barrier and blood-spinal cord barrier, and large molecules such as cytokines and neurotrophins have been shown to distribute faster to the spinal cord than to the brain after systemic administration (Pan and Kastin, 1999). We compared the effect of repeated systemic [Dmt¹]DALDA administration on [³H][Dmt¹]DALDA binding in the brain and spinal cord and found a similar reduction in binding (~30-35%) in both brain and spinal cord, suggesting comparable distribution of [Dmt¹]DALDA to the two sites. This small reduction in receptor number is unlikely to account for the development of tolerance, especially when similar loss of receptor numbers in brain and spinal cord was associated with a far greater loss of potency in the spinal cord.

In addition to the rapid onset of tolerance after systemic [Dmt¹]DALDA administration, our results also revealed cross-tolerance to morphine in these mice, with the s.c. morphine dose-response curve being shifted 15-fold. The finding of complete cross-tolerance to morphine after five doses of [Dmt¹]DALDA suggests the likelihood that they are acting via the same mechanism of action. Binding data with [³H][Dmt¹]DALDA are consistent with [Dmt¹]DALDA and morphine acting at the same µ-receptor. The K_d for [³H][Dmt¹]DALDA is similar to the K_i reported for [Dmt¹]DALDA against [³H]DAMGO binding (Schiller et al., 2000). The binding of [³H][Dmt¹]DALDA to brain membranes was completely displaced by morphine and DAMGO. Furthermore, the K_i for DAMGO is similar to the K_d reported for [³H]DAMGO in mouse brain, and the B_max determined with [³H][Dmt¹]DALDA is comparable with the B_max reported for [³H]DAMGO (Bhargava et al., 1997).

Our finding of cross-tolerance to morphine in [Dmt¹]DALDA-treated mice is in contrast to the report of lack of cross-tolerance when [Dmt¹]DALDA was given to mice after repeated morphine exposure (Neilan et al., 2001). In that study, mice were treated with morphine twice a day for 4 days using an escalating dose paradigm that only resulted in a 3-fold shift in the morphine dose-response curve. The relatively low level of tolerance was most likely due to the inadequate dosing frequency because the duration of action of morphine is only ~3 h. This level of tolerance may be too low for significant cross-tolerance, especially for a drug that is as potent and long-acting as [Dmt¹]DALDA. We have found that s.c. injection of morphine (10 times ED₅₀, every 12 h) for 2.5 days only resulted in a 2-fold shift in the dose-response curve for both morphine and [Dmt¹]DALDA (data not shown). Thus, the lack of cross-tolerance to [Dmt¹]DALDA in morphine-treated animals may, at least in part, be due to inadequate morphine exposure. Another possibility is that even if [Dmt¹]DALDA and morphine bind to the same receptor, the two ligands may have different intrinsic efficacies. Depending on the system studied, morphine has been found to be a partial agonist or very close to a full agonist in activation of GTPS binding (Emmerson et al., 1996; Remmers et al., 2000). [Dmt¹]DALDA behaved as a full agonist in the guinea pig ileum (Schiller et al., 2000), and its efficacy was determined to be 90% compared with DAMGO in activation of GTPS binding at the human µ-opioid receptor expressed in Chinese hamster ovary cells (G.-M. Zhao and H. H. Szeto, unpublished results). Thus, differences in intrinsic efficacy are not likely to be sufficient to account for the apparent lack of bi-directional cross-tolerance. It has recently been proposed that endocytosis of the receptor may reduce the development of opioid tolerance (Finn and Whistler, 2001). The ability of [Dmt¹]DALDA to cause internationalization of the µ-opioid receptor has not been studied, but morphine was reported not to cause appreciable internalization of the µ-opioid receptor (Sternini et al., 1996). It is, therefore, unlikely that differences in ability to cause receptor endocytosis can explain the lack of bi-directional cross-tolerance between [Dmt¹]DALDA and morphine.

Considering the very high µ/ selectivity of [Dmt¹]DALDA, it was surprising that concurrent administration of NTB with repeated [Dmt¹]DALDA dosing significantly reduced the magnitude of tolerance to [Dmt¹]DALDA by 50%. Previous studies have shown concurrent administration of -antagonists reducing the extent of tolerance to morphine, but morphine can also activate -receptors. The dose of NTB had no effect on [Dmt¹]DALDA potency in naive mice, whereas it significantly increased the ED₅₀ of deltorphin, confirming that [Dmt¹]DALDA was acting primarily at µ-receptors. It is, therefore, unlikely that activation of -receptors by [Dmt¹]DALDA initiated the cellular processes leading to tolerance. However, -receptors may still be involved in the regulation of µ-opioid receptor function after repeated dosing. In [Dmt¹]DALDA-tolerant mice, the addition of NTB significantly shifted the dose-response curve of i.t. [Dmt¹]DALDA to the same extent as if NTB was coadministered with [Dmt¹]DALDA all along. This ability of a -antagonist to enhance the potency of i.t. [Dmt¹]DALDA in tolerant animals was confirmed with a more selective -antagonist, TIPP[]. TIPP[] was given simultaneously with [Dmt¹]DALDA i.t. because it is not systemically active. The extent of reversal of tolerance by TIPP[] was dose-dependent and could be observed with as little as 0.2 nmol of TIPP[]. At a dose of 2 nmol, the magnitude of tolerance was reduced 4-fold.

The studies reported here do not reveal the mechanism behind the interaction between - and µ-receptors in the tolerant state. Complex interactions between the binding of µ- and -ligands have led to the suggestion of a physical association between the two receptor subtypes (Rothman et al., 1983). One possible mechanism is a physical association as in µ--heterodimers described in cell cultures coexpressing µ- and -receptors (George et al., 2000; Gomes et al., 2000). Interestingly, the addition of either a -agonist or -antagonist was reported to double the number of µ-binding sites (Gomes et al., 2000). The authors proposed that heterodimerization between µ- and -receptors alters the binding pocket of both receptors and that the binding of one ligand can restore the binding site of the other. It is possible that repeated exposure to a µ-agonist may lead to formation of µ--heterodimers that effectively results in a reduction of measurable binding sites, and the addition of a -ligand can restore receptor number. In our experiments, however, the addition of TIPP[] did not increase [³H][Dmt¹]DALDA binding in spinal cords from tolerant mice. The implication of µ--heterodimers in tolerance is uncertain, and the existence of µ- dimers has not been demonstrated in vivo. Interestingly, electron microscopic studies have shown that the -receptor is primarily localized intracellularly in the mammalian CNS rather than on the membrane (Arvidsson et al., 1995; Cheng et al., 1995). However, chronic pretreatment of neurons with morphine resulted in the targeting of the -receptor from intracellular vesicles to the membrane surface (Cahill et al., 2001). This would suggest that a physical interaction between µ and -receptors may be possible in the tolerant state.

A second possibility is that repeated exposure to [Dmt¹]DALDA results in up-regulation of an endogenous opioid peptide system, such as enkephalins, and activation of -receptors by this system may lead to inhibition of µ-receptor function. It was recently reported that i.t. administration of Leu-enkephalin (LE) inhibits the analgesic response to i.t. morphine, and this antianalgesic action of LE was blocked by naltriben, a ₂-antagonist (Rady et al., 2001). This is opposite to the synergistic action that LE has on morphine analgesia (Vanderah et al., 1996). The antianalgesic action of LE is observed with doses that are 1000-fold lower than those needed for enhancing morphine action. Interestingly, we found that a very small dose of TIPP[] (2 nmol/mouse) was able to reverse the tolerance to [Dmt¹]DALDA in the spinal cord. This dose of TIPP[] was unable to antagonize an ED₈₀ dose of i.t. DELT. It remains to be determined whether spinal tolerance is associated with elevated release of LE that may then serve to inhibit the action of opioid agonists, thus, effectively being perceived as tolerance. Under these circumstances, the administration of a -antagonist could reduce tolerance by removing the inhibitory action of endogenous LE. The recruitment of the -receptor to the surface after repeated treatment with a µ-agonist would be consistent with such a proposed model.

In summary, our results suggest that agonist activation of the -receptor is not required for development of opioid tolerance; however, -receptors play an important modulatory role in the maintenance of the tolerant state. Highly selective -antagonists such as TIPP[] or mixed µ-agonist/-antagonists may be useful in restoring the sensitivity to µ-agonists in the tolerant individual. The mixed µ-agonist/-antagonist, DIPP-NH₂[], showed less tolerance after i.c.v. administration compared with morphine (Schiller et al., 1999). Repeated intraperitoneal injection of a new nonpeptidic µ-agonist/-antagonist (SoRI 9409) apparently did not result in tolerance (Wells et al., 2001). However, this compound was only active in the writhing test after intraperitoneal administration, and recent studies showed that it failed to stimulate [³⁵S]GTPS binding at cloned opioid receptors (Xu et al., 2001). The issue of spinal tolerance has not been examined with these mixed µ-agonist/-antagonists. It will be very interesting to see if spinal tolerance will be limited with repeated use of these compounds.