Inhibition of Morphine Tolerance Development by a Substance P-Opioid Peptide Chimera1

Materials and Methods

Drugs and Peptide Synthesis.

Naltrexone (NTX) and MS were generous gifts from the National Institute on Drug Abuse, whereas RP67580 was a generous gift from Rhone-Poulenc Rorer (France). Radiolabeled materials were purchased from DuPont-New England Nuclear (Boston, MA), whereas all other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO), unless otherwise indicated. ESP6 was synthesized in our laboratory by the solution method published previously (Lipkowski et al., 1981), purified by standard HPLC procedures, and presented as a peptide-cyclodextrin complex, molar ratio 1:2, for facilitated delivery in vivo. As previously described for ESP7/β-cyclodextrin (βCD) complex, the βCD vehicle was found to be without analgesic effects when administered alone (Foran et al., 2000).

Intrathecal Catheterization.

All experimental procedures used in the present study were approved by the Tufts University Animal Research Committee, protocol 05-97. Adult male Sprague-Dawley rats (200–250 g) were implanted with chronic indwelling intrathecal catheters using a modification of the original method of Yaksh and Rudy (1976). Silastic tubing catheters (i.d. = 0.30 mm; o.d. = 0.64 mm, 11.5 cm in length) were inserted 7.5 cm into the intrathecal space to the level of T13-L1. After implantation of catheters, rats were allowed to recover from the surgery for 3 to 4 days. Animals exhibiting any sign of neurological or motor impairment, as evidenced by paralysis, abnormal gait, weight loss, or negligent grooming, were excluded from the study. Rats were housed separately to ensure catheter patency in a temperature- and light-controlled environment, with free access to food and water. After completion of drug testing, the catheter position was verified in each animal by postmortem examination of the spinal cord. ESP6-β-cyclodextrin complex and MS were dissolved in sterile saline and injected in a volume of 10 μl, followed by 10 μl of saline to flush the catheter (dead volume of 10 μl).

Tail-Flick Assay.

During recovery from surgery, the rats were habituated to the laboratory environment and to the analgesic testing apparatus. For measurement of the thermal antinociceptive properties of ESP6, a custom-made tail-flick apparatus consisting of a variable-intensity 300-W quartz projector bulb and a photodetector-automatic timer sensitive to 0.01-s intervals, was used. During testing, rats were placed in the tail-flick chamber, a light source was directed at the underside of their tail, and the latency to remove the tail was recorded. The baseline latency was approximately 3.5 s and the cutoff latency was set to 10 s (approximately 3 times baseline latency) to avoid tissue damage. Three measurements were made at each pre- and post-treatment time point and the results were averaged. Responses were expressed as percentage maximum possible effect (%MPE): [%MPE = ((post-treatment latency − baseline latency)/(cutoff latency − baseline latency)) × 100]. The area under the curve (AUC) was calculated for each dose on each day using the trapezoidal method (Tallarida and Murray, 1981). A baseline of zero was used to calculate both AUC and statistical significance. The data were evaluated with one-way repeated measures ANOVA followed by Bonferroni-corrected pairwise comparisons. Dunnett's test was used for all pairwise multiple comparisons versus control. Significance was defined as a P value <.05. It was expected that four animals would permit four Bonferroni-corrected contrasts to be made at an overall power of at least 80%.

Receptor Binding Assays.

Competitive binding analyses, performed to evaluate the relative affinities of ESP6 for the MOR and NK-1Rs, used rat brain membranes, prepared as a modification of published methods (Charlton and Helke, 1985; Zadina et al., 1994). Briefly, for MOR binding, fresh frozen rat brains were homogenized in 40 volumes of standard buffer (50 mM Tris-HCl (pH 7.4), 0.2 mg/ml BSA, 2.5 mM EDTA, 40 μg/ml bacitracin, 30 μg/ml bestatin, and 5 mM MgCl₂). After centrifugation at 15,000g for 20 min, the pellet was washed with standard buffer (+100 mM NaCl), followed by standard buffer alone. The membrane preparation was resuspended in 10 volumes of incubation buffer (standard buffer with 4 μg/ml leupeptin and 2 μg/ml chymostatin). NK-1R binding analyses used the same tissue preparation procedure, with the elimination of the NaCl wash and substitution of 3 mM MnCl₂ for 5 mM MgCl₂. Competitive MOR binding assays were performed at 4°C for 90 min, with each tube containing 1 mg of brain membrane suspension, 1.85 nM [³H]DAMGO, and increasing concentrations of nonradiolabeled competitor (DAMGO, ESP6). All assay concentrations were run in triplicate and each competitive binding analysis was repeated twice. For MOR binding, nonspecific binding was quantified as retained radioactivity in the presence of a 10,000-fold excess of nonradioactive displacer, 10 μM DAMGO (gift from National Institute on Drug Abuse). After incubation, bound radioactivity was separated from free ligand via rapid filtration through GF/B glass fiber discs, preadsorbed in 0.5% polyethyleneimine and 50 mM Tris-HCl (pH 7.4) to minimize nonspecific binding, using a Brandell-Harvester apparatus. Competitive NK-1R binding analyses were performed exactly as described for the MOR binding analyses, with the exception that incubation was performed at room temperature for 75 min. The radioligand was an¹²⁵I-labeled SP-Bolton Hunter analog added at a final concentration of 0.1 nM. Nonspecific binding was quantified as retained radioactivity in the presence of a 10,000-fold excess of nonradioactive displacer, 1 μM SP. All binding data were analyzed using GraphPad PRISM and fit to a sigmoidal curve using nonlinear regression. K_i values for ESP6 at each receptor were calculated.

Measurement of Inositol Phosphate (IP) Formation.

The NK-1R is coupled to a stimulatory G_q protein. Activation of the NK-1R increases the levels of phospholipase C, which subsequently cleaves phosphatidyl inositol into inositol triphosphate and diacylgycerol. As a complementary measure of NK-1R activation, the potency of ESP6 to simulate IP production was assessed. Following the protocol of Blaker et al. (1998), 10⁶ COS-7 cells/10-cm plate were transfected with 5 μg of rat NK-1R receptor cDNA (Kage et al., 1996) or control pcDNA1.1 plasmid (Invitrogen, San Diego, CA). After transfection, cells were split into 12-well plates (2 × 10⁵ cells/well) and then labeled overnight with 3 μCi/ml of the inositol phosphate precursor myo-[³H]inositol. The following day, cells were stimulated for 60 min with ESP6 (0.01–12,000 nM) or SP (1,000 nM) in the presence of 10 mM LiCl (duplicate samples/reaction). LiCl was included to inhibit the degradation of inositol phosphates. After extraction with methanol/chloroform to remove inositol metabolites in the organic phase, IPs in the aqueous phase were separated from other acidic tritiated products by strong anion exchange chromatography. Retained IPs were eluted with 2 M ammonium formate. To minimize intra-assay variability, IP production is expressed as a ratio of tritiated IPs/total tritium incorporated (i.e., as a fraction of the total cellular tritium content that was incorporated into cells during overnight exposure to the IP precursor myo-[³H]inositol). The EC₅₀ for ESP6 was calculated by nonlinear regression analysis using GraphPad PRISM.

Measurement of ESP6-Mediated Inhibition of cAMP Production.

HEK293 cells in 24-well (5 × 10⁴cells/well) or 96-well (1 × 10⁴ cells/well) plates were transiently transfected with 200 ng/ml of wild-type rat MOR and 1000 ng/ml of a cAMP-responsive reporter gene construct using a lipofection method (Felgner et al., 1987). The reporter gene encoded firefly luciferase under the control of a cAMP-responsive element (CRE), and was a generous gift of Dr. M. R. Montminy (Montminy et al., 1996). Control cells were transfected with 200 ng/ml of control pcDNA1.1 plasmid and 1000 ng/ml of the CRE-luciferase construct to assess MOR-independent reporter gene activity.

To assess MOR-mediated inhibition of cAMP-dependent luciferase activity, experiments were performed in the presence of 10 μM forskolin, which activates adenylate cyclase. To define a concentration-response curve for DAMGO-, EM-2-, and ESP6-induced inhibition of adenylyl cyclase/cAMP activity, HEK293 cells were concomitantly treated with a range of concentrations of these peptides (0.01–10,000 nM). Due to the lag-time in the detection of cAMP-induced luciferase transcription and subsequent protein activity, longer stimulation periods were used compared with assays, which directly measure cAMP production. The optimal time course of ligand-induced inhibition of cAMP was explored. Based on these experiments, a 6-h time point at 37°C in serum-free media was used for further study (triplicate samples/reaction). For experiments completed in 24-well plates, cells were subsequently lysed in extraction buffer containing 1% Triton X-100. Light emission was measured immediately using a luminometer (Monolight 2010). Cell lysate from transfected cells (50 μl) was mixed with 250 μl of ATP buffer (43 mM glycylglycine, pH 7.8, 22 mM MgSO₄, 0.4 mg/ml BSA, 0.6 mM EDTA, pH 8.0, 1 M dithithreitol, and 5 mM adenosine 5′-triphosphate). The solution was placed in the luminometer and the bioluminescent reaction was initiated by injecting 100 μl of d-luciferin substrate. Light emission was measured during a 20-s period. For experiments in 96-well plates, luciferase activity was measured using the LucLite assay kit (Packard, Meriden, CT) and a Packard microplate scintillation counter. The IC₅₀ values of DAMGO, EM-2, and ESP6 (inhibition of forskolin-induced luciferase activity) were calculated by nonlinear regression using the GraphPad PRISM computer program.

Different cell lines were used to study signaling of ESP6 via NK-1R and MOR. COS-7 cells are well established as hosts with which to monitor signaling through G_q-coupled receptors (i.e., NK-1R). However, in our hands COS-7 cells are not as effective at determining signaling through G_i-coupled receptors (i.e., MOR). The reason for this may be that, at least in transient assays, only a fraction of the cells express MORs (G_i coupled), whereas all cells are uniformly stimulated by forskolin (stimulation of adenylate cyclase). We have previously established a system for monitoring luciferase activity in HEK293 cells, which works well. This provided an opportunity to solve the discrepancy between receptor expression and forskolin stimulation. When the CRE-luciferase construct and MOR are coexpressed in HEK293 cells the stimulatory effects will only occur in cells with the inhibitory MOR present. For this reason, HEK293 cells were used to study signaling of ESP6 via MOR. Because similar experiments can be performed with each peptide chimera, theK_i and IC₅₀ results can be used to make relative comparisons about peptide potency and affinity. Additional in vivo and in vitro data can support these comparisons.

Results

Intrathecal Administration of ESP6 in Combination with MS.

The pharmacological interactions of the EM-2/SP chimera, designated ESP6 (Fig. 1), and the prototypic opioid analgesic MS were evaluated. ESP6 administered intrathecally at 1.0 μg/animal produced an analgesic response, which plateaued at approximately 10% MPE at early time points and was not significantly different from baseline (Fig. 2A). Similar results were seen with 0.05 and 0.1 μg of ESP6 alone (data not shown). In contrast, ESP6 at 0.05, 0.1, or 1.0 μg coadministered with 0.2 μg of MS or 0.2 μg of MS alone, effected significant analgesic responses for 30 to 45 min, as measured by the tail-flick test (Fig. 2A). Analgesic responses were observed to plateau at 30 to 60% of MPE. Tolerance development to these same dosages of ESP6 in combination with MS, 1.0 μg of ESP6 alone or 0.2 μg of MS alone, was monitored over a 5-day course of administration (Fig. 2B). Importantly, 0.05 and 0.1 μg of ESP6 with MS, given once daily, effected equivalent analgesic responses over time. Although the level of analgesia produced by 0.05 and 0.1 μg of ESP6 with MS was different depending on the dose, tolerance did not develop to either dose over a 5-day period. At the 1.0-μg dose of ESP6 with MS, tolerance developed by day 4. By comparison, a decline in analgesic potency of 0.2 μg of MS administered alone was observed with at_1/2 of approximately 1 day and complete tolerance or a return to baseline was seen by day 3. ESP6 (1.0 μg) alone did not produce a response significantly different from baseline on any day of treatment. The vehicle βCD alone had no effect on analgesia.

Analgesia Produced by ESP6 in Combination with MS: Blockade by Naltrexone.

The opioid dependence of the potentiated analgesic response effected by 0.1 μg of ESP6 in combination with 0.2 μg of MS was evaluated using the opioid antagonist NTX (Fig.3). On day 1, the combination of agents was observed to produce a strong analgesic response. On day 2, rats were pretreated with 2.0 μg of NTX by the intrathecal route 10 min before ESP6 and MS administration. The predicted analgesic response produced by ESP6 and MS was completely blocked by NTX action, indicating opioid receptor mediation of the observed pharmacological phenomenon. On day 3, NTX was not administered before ESP6 and MS and an analgesic response similar to day 1 resulted. NTX alone did not produce of response significantly different from baseline. Similar results were seen with 1.0 μg of ESP6 in combination with 0.2 μg of MS and NTX (data not shown).

Inhibition of MS Tolerance Development by ESP6: Reversibility by the NK-1R Antagonist RP67580.

The ability of ESP6 in combination with MS to effect an analgesic response with delayed opioid tolerance was apparently mediated by NK-1Rs, based on reversibility by the selective NK-1R antagonist RP67580 (Fig.4). Rats were pretreated with 250 pmol of RP67580 by the intrathecal route 10 min before 0.1 μg of ESP6 and 0.2 μg of MS on days 1 to 4. ESP6 and MS displayed at_1/2 of approximately 1 day for the decay of opioid efficacy as a function of time, similar to MS alone. The analgesia produced by ESP6 and MS was only significantly different from the analgesia produced by MS alone on day 1. On day 5, RP67580 was not given and no analgesic effect was observed. RP67580 alone produced no analgesic effects. Similar results were seen with 1.0 μg of ESP6 in combination with 0.2 μg of MS and RP67580 (data not shown).

Binding Affinities and Activation Potencies of ESP6 at the MOR and NK-1Rs.

The ability of ESP6 to interact with the both MOR and NK-1Rs in vivo was confirmed by in vitro binding assays using rat brain membranes (Table 1). ESP6 displayed aK_i of 92 nM (95% CI = 43–198 nM;n = 2) for the inhibition of binding of radiolabeled DAMGO to the MOR (data not shown) and a K_iof 305 nM (95% CI = 41–2263 nM; n = 2) for the inhibition of binding of radiolabeled SP (data not shown). ESP6 is selective for the MOR over the δ- and κ-opioid receptors and for the NK-1R over the NK-2 and NK-3 receptors. Similar results were obtained with ESP7 (Foran et al., 2000). Additional in vitro assays with recombinant receptors were performed to demonstrate the ability of ESP6 to act as an agonist at the NK-1R or the MOR. The potency of ESP6 at the rat NK-1R was confirmed using the recombinant rat NK-1R transiently expressed in COS-7 cells. The capacity of ESP6 to stimulate IP production was quantified using a range of peptide concentrations. ESP6 possessed an EC₅₀ of 325 nM for IP formation (Fig. 5A), which is two to three orders of magnitude higher than the reported value for SP (Nielsen et al., 1998). IP production by a saturating concentration of SP (1 μM) was compared with the highest concentration of ESP6 (12 μM) and a significant difference in stimulation was observed. Neither ESP6 nor SP evoked IP production in cells transfected with the empty expression vector pcDNA1.1 (data not shown).

The potency of activation of ESP6 at the rat MOR was determined in HEK293 cells simultaneously transfected with the rat MOR and a CRE-luciferase construct. DAMGO, EM-2, and ESP6 inhibited forskolin-stimulated activity with IC₅₀ values of 8.0 nM, 4.8 nM, and 1.9 μM, respectively (Fig. 5B). NTX blocked the ability of DAMGO, EM-2, and ESP6 to inhibit forskolin-induced activity, whereas NTX alone had no effect on forskolin-induced activity (data not shown). Finally, ESP6 was observed to mediate no pharmacological effect in cells transfected with the empty expression vector pcDNA1.1 and CRE-luciferase alone (data not shown).

Discussion

The salient results of our study are as follows. Intrathecal administration of ESP6 does not directly effect analgesic responses. Daily intrathecal administration of varying concentrations of ESP6 in combination with a modestly effective concentration of MS, however, results in prolonged analgesic responses compared with those produced by MS alone. The analgesia produced by ESP6 and MS is opioid in nature, based on the ability of the opioid receptor antagonist NTX to completely block the pharmacological response. Unlike MS alone, the analgesia produced by ESP6 and MS is sustained over a 5-day time course, indicating the ability of ESP6 to significantly inhibit tolerance development. When ESP6 and MS are coadministered with a high-affinity NK-1R antagonist, RP67580, a decay in analgesic potency similar to that seen with MS alone is observed. This suggests that coincident activation of spinal NK-1Rs is necessary to delay tolerance. Confirmatory in vitro biochemical analyses indicate that ESP6 acts independently as an agonist at the both MOR and NK-1Rs, albeit with considerably lower affinity than previously demonstrated for the parent compounds EM-2 and SP (Table 1).

Comparison of the EM-2 and SP amino acid sequences revealed two phenylalanine residues, which were shared by both peptides, thereby representing a naturally occurring, flexible hinge region capable of separating the distinct functional domains within the chimera. This requisite overlapping diphenylalanine hinge region was originally used in conjunction with the NH₂- and COOH-terminal domains of EM-2 and SP, respectively, to create the first chimeric peptide, ESP7, capable of significant opioid analgesia with minimal tolerance development (Foran et al., 2000). Similar to ESP7, ESP6 contains overlapping NH₂- and COOH-terminal domains of EM-2 and SP, respectively, with a proline at position 5 to confer enhanced stability and selectivity for the NK-1R (Laufer et al., 1986). The observed differences in analgesic profiles between ESP6 and ESP7 may therefore reside in the extent of conformational flexibility of their respective opioid domains. Despite the absence of direct analgesic action, our present results indicate the ability of ESP6 to independently bind to the MOR and NK-1Rs, and activate secondary messenger pathways through both the MOR and NK-1Rs expressed in cultured cells. ESP6 possesses K_i values of 305 and 92 nM for binding to the NK-1R and MORs, respectively, an EC₅₀ of 325 nM for NK-1R-mediated stimulation of IP production, and an IC₅₀ of 1900 nM for MOR-mediated inhibition of adenylate cyclase activity. The disparity of over an order of magnitude in binding affinity versus activation potency at the MOR emphasizes the effect of a proline at position 5 and suggests that ESP6 may possess partial opioid agonist/antagonist properties. This may explain the lack of direct analgesic action of the chimera. Our findings suggest that, at least at the tested doses, the pharmacological effectiveness of ESP6, in contrast to its analog ESP7, relies primarily on its interaction with the NK-1R. In comparison to ESP7, conformational constraints in its derivative ESP6 appear to prevent significant contribution of MOR activation in the maintenance of opioid analgesia over time.

In rodent models, MS analgesia decays rapidly (t_1/2 = 1 day) (Kream et al., 1993). Similarly, in this manuscript a t_1/2 = 1 day for MS analgesia was obtained using intrathecal delivery to male Sprague-Dawley rats and the tail-flick assay. Intrathecal administration of three nonanalgesic doses of ESP6 with MS leads to a prolongation of MS analgesia in the rat tail-flick test. When the same dose of ESP6 is delivered once a day for five days with MS, MS responsiveness is maintained longer in rats given ESP6 than in rats given MS alone. ESP6 (0.05 and 0.1 μg) delay tolerance development to MS for at least 5 days, whereas the highest dose of ESP6 (1.0 μg) delays tolerance for 4 days. Although 0.05 μg of ESP6 appears to decrease the analgesic efficacy of MS and 0.1 and 1.0 μg of ESP6 appear to increase the analgesic efficacy of MS as seen in Fig. 2B, there is no significant difference on day 1 between the analgesia produced by MS alone and MS with any dose of ESP6. The lack of an apparent dose-dependent effect by the tested doses of ESP6 may be due to its ability to concurrently bind to two opposing spinal receptor systems. Furthermore, the affinity and potency of ESP6 for the MOR and NK-1Rs may offer some explanation for its pharmacological actions. Previous studies using ESP7 suggest that this chimera, with roughly the same affinity and activation potency at the MOR (218 and 95 nM, respectively), is capable of effecting significant opioid analgesia (Table 1). ESP6 displays an order of magnitude lower activation potency at the MOR (1900 versus 95 nM for ESP7), thereby representing a potential opioid adjuvant, not a direct-acting analgesic. Although ESP6 does not activate the MOR potently enough to effect analgesia alone, all doses of ESP6 given in this article delay tolerance to MS. More convincing results are seen with 0.05 and 0.1 μg of ESP6. The ability of the NK-1R antagonist RP67580 to reverse the effects of ESP6 on MS tolerance suggests that ESP6s potency (i.e., EC₅₀= 325 nM) at the NK-1R is sufficient to stimulate the receptor and slow tolerance development. The decline in analgesic potency after 3 days, with 1.0 μg of ESP6, may be due to direct mixed agonist/antagonist activity at the MOR, compared with 0.05 and 0.1 μg of ESP6, and not to lack of efficacy at the NK-1R. Alternatively, the inability of ESP6 to stimulate IP production to a similar level as a saturating concentration of SP indicates that either ESP6 is a partial agonist at the NK-1R or more likely that a higher dose of ESP6 is needed to reach maximal stimulation. Activity at the NK-1R or MOR appears to mediate separate distinct effects (e.g., tolerance development or analgesia).

The results with ESP6, in conjunction with our previous data on ESP7, provide valuable insight on the ability of NK-1R activation to modulate opioid analgesia and the effect of a peptide structure and function on analgesic effectiveness (Table 1). Our previous studies using ESP7 indicate a required equivalence of potencies (similar IC₅₀ values and EC₅₀ values at both receptor sites) to effect sustained opioid analgesia over time. In the case of ESP6, it appears that the proline-induced conformational restraint following the hinge region both diminishes the affinity of the opioid moiety to activate the MOR and confers partial antagonist properties to the chimera. In effect, the lack of equivalence in intrinsic properties of receptor recognition domains markedly diminishes direct analgesic properties of the peptide. These empirical rules of construction will apply to future design of opioid/SP peptide chimeras with the caveat that retention of activity at the NK-1R appears to be crucial to maintaining analgesic responsiveness over time.

As discussed previously with regard to ESP7, tolerance can be delayed under certain circumstances. Notably, rats receiving nociceptive stimulation have delayed tolerance to MS analgesia compared with naı̈ve rats (Colpaert et al., 1978). The extent of opioid tolerance seen in patients also depends on the type of pain present. As an example, patients with postoperative pain demonstrate markedly attenuated tolerance development (Carr et al., 1998). These preclinical and clinical results suggest a beneficial role for excitatory neuropeptides, such as SP, in hindering tolerance to opioid analgesia. The need to optimize analgesic endpoints and the potential interaction of SP and opioids in tolerance development prompted the design and synthesis of chimeric peptides with opioid and SP moieties. The peptides can be used to understand spinal nociceptive processing and to compare the structure/function relationships of peptides with slightly different amino acid sequences. These peptides have the potential to act as pain relievers alone or as adjuvants to traditional opioids analgesics. The peptide discussed in this article acts at both the MOR and NK-1Rs to enhance MS analgesia after daily and repeated opioid administration. ESP6 represents a valuable probe to monitor the interactions of tachykinin/opioid systems, as well as a potential adjuvant therapy to treat acute and chronic pain.