Design of Analgesic Trivalent Peptides with Low Withdrawal Symptoms: Probing the Antinociceptive Profile of Novel Linear and Cyclic Peptides as Opioid Pan Ligands

Abstract

The discovery of efficacious and safe analgesics with reduced side effects is the foremost challenge in the pain field. In this work, we report the in vitro and in vivo evaluation of linear and cyclic analogues of biphalin with the aim to complete the series of structural modifications previously applied in the development of opioid peptides incorporating a xylene bridge. Replacement of Tyr^1,1′ by Dmt (2,5-dimethyltyrosine) in the linear biphalin analogue AM94 and cyclic analogue MACE4 resulted in two new compounds (namely, MJ2 and MJ5) endowed with improved KOR/MOR/DOR binding affinity. Both compounds showed a strong antinociceptive profile in in vivo models of nociception, allodynia, and hyperalgesia via the tail flick, hot plate, and formalin tests after intracerebroventricular and subcutaneous administration. One of these ligands, MJ2, was also tested in tolerance and dependence studies, exhibiting very little withdrawal symptoms.

Graphical Abstract

INTRODUCTION

Opioids are powerful drugs for the treatment of moderate to severe pain, but their use is associated with important side effects such as respiratory depression, addiction, sedation, and constipation. Their clinical application is unchallenged in acute and cancer pain, but inappropriate use in chronic non-malignant pain has contributed to the opioid crisis, mainly observed in some Anglo-Saxon countries.^1,2 Therefore, new analgesics with reduced side effects are badly needed, and opioids still appear to be the most promising of current approaches to analgesic development. Some biased, multivalent, and peripherally selective agonists have moved on to phase III studies, but few new drugs are currently available for clinical applications.^3,4

One novel approach is multifunctional ligands with activity at multiple opioid receptors.⁵ However, even compounds endowed with high binding affinity for μ-opioid receptor (MOR) and mixed MOR/δ-opioid receptor (DOR) agonists demonstrate unwanted side effects which render them poorly suitable as drug candidates^.5,6 Activation of DOR produces spinally mediated antinociception that is enabled by inflammation, which is generally less pronounced than that produced by MOR stimulation;⁶ DOR antagonists may also prevent MOR-mediated tolerance development and dependence/withdrawal.⁷ The phenomena of reward/reinforcement involve selective DOR agonists with which the presence of MORs is also required to generate reinforcing properties.^8–10

K-opioid receptor (KOR) agonists have been shown to produce pain relief and can reduce drug-seeking; however, they can also provoke dysphoria and sedation.^11,12 A full/partial MOR agonist with KOR agonist activity (e.g., pentazocine, nalbuphine, and butorphanol) shows a decreased abuse liability, even though some patients have reported dysphoria and withdrawal was detected in opioid-addicted patients.¹³ The development of a partial KOR agonist endowed with MOR activity could be a valuable alternative to reach non-addictive analgesic activity without dysphoria, aiming to mitigate and titrate rewarding versus dysphoric behavior of the compound.^14–16 Together, these studies suggest that an opioid ligand with agonist activity at all three receptors could synergize the antinociception common to each receptor while reducing side effects which are antagonistically produced by each receptor (e.g., the anti-reward effects of KOR balancing the pro-reward effects of MOR).

Recently, Journigan et al. reported a close structural analogue of the KOR-selective antagonist JDTic, namely, AT-076; this compound is a pan antagonist with nanomolar affinity for all three opioid receptors.¹⁷ A structure–activity relationship (SAR) has been delineated by the authors suggesting the presence of a possible chemotype able to bind with high affinity at the three opioid receptors, acting as an universal opioid scaffold. In contrast to the trivalent partial agonists described by Toll et al. (e.g., PPL-103,101) which also deserve consideration as potent antinociceptive compounds with a lack of aversion in mice,¹⁵ AT-076 is a non-morphinan opioid ligand. These observations oriented our investigation on structural modifications of the mixed MOR/DOR agonist AM94, a p-F-Phe^4,4′ analogue of biphalin incorporating a piperazine linker.¹⁸ Recently discovered cyclic peptides MACE1–4, which we derived from the AM94 scaffold, showed a broad range of K_I values and diverse efficacy/potency on MOR, DOR, and KOR.¹⁹ Compound MACE4, a cyclic analogue of biphalin incorporating the o-bis(methyl)benzene bridge, p-F-Phe^4,4′ and Dmt^1,1′, possesses a high binding affinity for MOR/DOR and a moderate affinity for KOR, resulting in a full agonist at MOR and a partial agonist at DOR/KOR, with the best efficacy and in vivo antinociceptive activity of the series.¹⁹ The insertion of a Dmt residue in place on the native Tyr present in the analogue MACE2 seems to be crucial to determine a well-defined orientation of the side chains in the three-dimensional structure that fits the MOR binding site, thus producing a subnanomolar activity at MOR and also an improvement in KOR affinity.^20,21

However, the increase in hydrophobicity through the two methyl groups on the phenolic ring in Dmt^1,1′ allowed further differentiation of the biological properties among the structurally related MACE1–4 compounds.¹⁹ Prompted by these findings, we planned to explore the influence of Dmt^1,1′ on the in vitro and in vivo antinociceptive activity of five linear and cyclic analogues of biphalin and AM94 (MJ1–5), thus completing our SAR study on these two lead compounds (Figure 1).

Figure 1.

New designed linear and cyclic analogues of biphalin (MJ1, 3 and MJ5) and AM94 (MJ2, 4).

Peptides MJ1, MJ3 and MJ5 are analogues of biphalin preserving the hydrazine linker, Phe^4,4′ and Gly^3,3′. The first is a linear peptide in which D-Ala^2,2′ is directly linked to Dmt^1,1′, while the last two are cyclic analogues containing a disulfide bridge between two D-Cys^2,2′ residues in peptide MJ3 or further functionalized with o-bis(methyl)benzene in peptide MJ5. Notably, this compound retains the Phe^4,4′ of biphalin, in contrast to the previously described analogue MACE4, where it is replaced by the p-F-Phe residue.¹⁹ Peptides MJ2 and MJ4 are linear and cyclic analogues of AM94, respectively, conserving the piperazine linker, p-F-Phe^4,4′ and Gly^3,3′ residues. In the first of them, D-Ala^2,2′ is also retained, instead of MJ4, where it is replaced by D-Cys^2,2′ to promote the formation of the o-bis(methyl)benzene bridge through the disulfide bond-containing intermediate.²² The combination of all these structural features could deeply influence the pharmacological profile of the new compounds in terms of binding affinity and analgesic efficacy on the three opioid receptors following different administration routes. Some of them behave as trivalent ligands with full agonist/partial agonist activity and a strong long-lasting antinociceptive profile after subcutaneous (s.c.) administration.

RESULTS AND DISCUSSION

Chemistry.

The peptide derivatives MJ1–5 were prepared following a well-established solution-phase peptide synthesis previously described.^19,23 Commercially available Fmoc-Dmt-OH and Boc-protected amino acids were employed as synthetic precursors. Standard coupling reactions were performed to obtain the complete amino acid sequence. Full deprotection under acidic conditions led to the desired peptides as TFA salts in good overall yields (analytical and structural characterizations are reported in the experimental section and the Supporting Information).

Evaluation of Compound Binding Affinity.

After synthesis, all compounds were characterized in vitro for their orthosteric binding affinity to the three opioid receptors, with the goal of identifying pan agonists. All compounds bound with high affinity to the DOR and with slightly less but still high affinity to the MOR (Figure 2). The compounds differentiated from each other in their affinity to the KOR. MJ3 and MJ4 bound with moderate affinity to the KOR, 470 and 230 nM, respectively, providing a modest selectivity of up to ~200 fold for MJ3 for KOR versus DOR at the highest. This profile was consistent with our earlier MACE series of ligands, which showed similar MOR/DOR over KOR selectivity.¹⁹ In contrast, MJ1, MJ2, and MJ5 showed a balanced profile of near-equal opioid affinity. MJ2 in particular had low-nM affinity at all three receptors and all within ~2 fold of each other. These results suggest that our strategy to generate pan ligands was successful.

Figure 2.

In Vitro binding affinity of MJ compounds. All compounds were competed against ³H-diprenorphine in binding assays against all three human opioid receptors expressed in CHO cells. Positive control compounds (naloxone for MOR and DOR, U50, 488 for KOR) were included to validate the assay. The concentration–response curves shown are summary curves with the mean value ± SEM from N ≥ 3 independent experiments for each receptor as noted. The K_I values in the table below were calculated separately for each independent experiment and reported as the mean ± SEM.

Evaluation of Compound Functional Activity.

The compounds were next evaluated for their functional activity using a ³⁵S-GTPγS coupling assay to determine if the compounds were balanced agonists as designed. First, all compounds displayed high potency and efficacy of full agonism at the MOR (Figure 3). There was a slight separation of potencies, with MJ1, 2, 5 showing ~0.2–1 nM potency and with MJ3–4 showing slightly lower potency at ~8–10 nM. This separation mirrored small affinity differences between the compounds observed in Figure 2. Notably, all compounds had higher MOR potency than affinity, suggesting they have high intrinsic efficacy at the MOR.

Figure 3.

In vitro functional activity of MJ compounds. All compounds were used to stimulate ³⁵S-GTPγS coupling downstream of all three human opioid receptors expressed in CHO cells. Positive control compounds (DAMGO for MOR, SNC80 for DOR, and U50, 488 for KOR) were included to validate the assay. The concentration–response curves shown are summary curves with the mean value ± SEM from N ≥ 3 independent experiments for each receptor as noted. The potency (EC₅₀) and efficacy (E_MAX) values in the table below were calculated separately for each independent experiment and reported as the mean ± SEM. The E_MAX of each compound was normalized to the stimulation caused by the positive control agonist, defined as 100% (with vehicle stimulation defined as 0%).

The picture became more complicated at the DOR. MJ1–2 and 4–5 showed high potency partial agonism, with efficacies from ~35 to 60% (Figure 3). MJ3 in contrast had no detectable agonist activity at the DOR. Considering the high affinity for this compound at the DOR in Figure 2, MJ3 may be a DOR antagonist. Overall, this profile of high potency DOR partial agonism was similar to our previous MACE series.¹⁹

Last, the compounds diverged from previous results most significantly at the KOR. The compounds showed a high range of potencies, from ~10 to 1000 nM, and also a high range of efficacies, from ~20 to 80% (Figure 3). These differences did not map neatly to the affinity differences at the KOR observed in Figure 2, suggesting that some compounds may have high (MJ5) and low (MJ3–4) intrinsic efficacy. MJ1 had very similar affinity and potency, suggesting balanced intrinsic efficacy. Unfortunately, an elevated baseline (BL) for MJ2 at KOR prevented full analysis, which sometimes occurs in our analyses,^19,24 although this compound had the highest efficacy. Overall, MJ1,2,4 had the highest potential as balanced MOR/DOR/KOR agonists, although complicated by the partial agonist activity at DOR and KOR. Given these findings, MJ2 represents the best balanced agonist with high potency, affinity, and efficacy at each opioid receptor.

In Vivo Experiments.

Hot Plate and Tail Flick Nociceptive Pain Tests.

In the first series of experiments, the response to thermal nociceptive stimuli was assessed after intracerebroventricular (i.c.v.) administration of MJ peptides at the dose of 0.6 nmol/10 μL using the hot plate and the tail flick test. The effects induced by MJ1–5 from 15 to 120 min after administration are reported in Figure 4. From 15 to 60 min after administration, MJ1, MJ2, and MJ5 induced a potent antinociceptive effect that was roughly similar to that observed after biphalin or the cyclic analogue of biphalin containing a disulfide bridge (BS-S) administration.²⁵ Interestingly, from 90 to 120 min after administration, MJ1, MJ2, and MJ5 maintain their antinociceptive effect, whereas biphalin and BS-S antinociceptive effects steadily decrease until they reach maximum possible effect (MPE) values similar to those observed in animals treated with vehicle (Figure 4). Both in the hot plate and in the tail flick test, MJ3 induced the same effects observed after biphalin or BS-S administration (Figure 4). Unlike all other peptides, MJ4 induced lower antinociceptive effects than those induced by biphalin and BS-S both in the hot plate and in the tail flick test. These results are somewhat surprising, in that both MJ3 and MJ4 display potent MOR agonist activity in Figure 3; their relatively poor performances here may be the results of stability/pharmacokinetic differences or similarities.

Figure 4.

Antinociceptive effects induced by MJ1–5 in the hot plate and tail flick test. MJ1–5, biphalinS-S (BS-S) or biphalin (B) was administered i.c.v. at the dose of 0.6 nmol/10 μL. V is for vehicle-treated animals (DMSO/saline 1:5 v/v, 10 μL/mouse). Data were reported as the time course of the percentage of MPE and statistically analyzed by two-way ANOVA, followed by Tukey’s multiple comparisons test. ^# is for p < 0.0.5, ^## is for p < 0.01, ^### is for p < 0.001, and ^#### is for p < 0.0001 versus V. N = 12/group.

Having observed that MJ1, MJ2, and MJ5 induced antinociceptive effects that did not decrease after 120 min of administration, we decided to investigate the duration of these effects by prolonging the observation time up to 240 min. The results of these experiments are reported in Figure 5. 150 min after administration, the antinociceptive effects of the peptides appear to decrease, even if the analysis of variance (ANOVA) showed no significant effects related to the observation time; however, significant antinociception remained after 2 full hours of treatment. These results suggest that MJ1,2 and MJ5 have very long durations of action, perhaps for pharmacodynamic or pharmacokinetic reasons, which is promising for potential translation to human therapies.

Figure 5.

Effects induced by MJ1, MJ2, and MJ5 administered i.c.v. at the dose of 0.6 nmol/10 μL in the hot plate and in the tail flick test. V is for vehicle-treated animals (DMSO/saline 1:5 v/v, 10 μL/mouse). Data were reported as the time course of the percentage of MPE and statistically analyzed by two-way ANOVA, followed by Tukey’s multiple comparisons test. ^# is for p < 0.0.5, ^## is for p < 0.01, ^### is for p < 0.001, and ^#### is for p < 0.0001 versus V. N = 7/group.

Formalin Test.

While promising, the hot plate/tail flick tests are of limited translational relevance. We thus next tested our series in the formalin model, a more relevant nociceptive and inflammatory pain state. In the formalin test, peptides were administered by s.c. route in the left dorsal surface of the hind paw at the dose of 100 nmol/20 μL, 15 min before formalin was injected in the same hind paw. The results of these experiments are reported in Figure 6A. After MJ1 administration, we observed a reduction of the licking behavior induced by formalin both in the early phase and in the late phase, but this effect did not reach statistical significance. After MJ2 and MJ5 administration, the peptides strongly reduced the nociceptive behavior induced by the formaldehyde in both the early and late phases, corresponding to nociceptive and inflammatory pain states, respectively. Peptide MJ3 was able to reduce the early and late phases, but statistical significance was reached in the late phase only, while MJ4 did not change the effects induced by formalin both in the early phase and in the late phase, mirroring our results in the hot plate and tail flick tests.

Figure 6.

Efficacy of MJ peptides to relieve formalin-induced pain. MJ1–5, biphalin (B) and biphalin S–S (BS-S) were administered (A) s.c. at the dose of 100 nmol/20 μL in the dorsal surface of the hind paw 15 min prior to formalin (1% in saline, 20 μL/paw) or (B) s.c. at the dose of 100 nmol/20 μL in the dorsal surface of the hind paw 1 h before formalin (1% in saline, 20 μL/paw). V is for vehicle-treated animals (DMSO/saline 1:5 v/v, 200 μL/mouse, s.c.). Data were reported as the total time the animal spent licking or biting its paw during the early and late phases of formalin treatment and were statistically analyzed by two-way ANOVA, followed by Tukey’s multiple comparisons test. ^# is for p < 0.0.5, ^## is for p < 0.01, ^### is for p < 0.001, and ^#### is for p < 0.0001 versus V. N = 7/group.

MJ2 and MJ5 showed the best pain relief, with long-lasting effects in tail flick and hot plate pain, as well as the highest efficacy in the formalin test, so we investigated their effects following s.c. administration in the left dorsal surface of the hind paw at the dose of 100 nmol/20 μL 1 h before formalin. The results of these experiments are reported in Figure 6B. Both MJ2 and MJ5 reduced the nociceptive effects of formalin in both early and late phases of the test, confirming their long-lasting efficacy in relieving pain after s.c. administration.

Zymosan-Induced Hyperalgesia and Allodynia.

The linear and cyclic peptides MJ2 and MJ5 induced the highest antinociceptive effects both in a thermal model and in an inflammatory model of pain; thus, we further investigated their effects in an animal model of hyperalgesia and allodynia. In these experiments, MJ2 and MJ5 were injected i.c.v. at the dose of 0.6 nmol/10 μL, 3.5 h after zymosan administration. The results of these experiments are reported in Figure 7. Compounds MJ2 and MJ5 induced the same antinociceptive effects as observed after biphalin or BS-S i.c.v. injection, 4 h after zymosan administration. When latencies to thermal or mechanical stimuli were recorded 6 h after zymosan administration, MJ2 and MJ5 demonstrated antinociceptive effects, whereas the two reference compounds were inactive (Figure 7). This suggests that MJ2 and MJ5 are superior analgesics to the biphalin and BS-S reference compounds.

Figure 7.

MJ2 and MJ5 relieve zymosan-induced hyperalgesia and allodynia. MJ2, MJ5 biphalin (B) or biphalin S-S (BS-S) was administered i.c.v. at the dose of 0.6 nmol/10 μL in zymosan-induced hyperalgesia and allodynia. V is for vehicle-treated animals (DMSO/saline 1:5 v/v, 10 μL/mouse). Data were evaluated as a percentage of the paw withdrawal latencies recorded prior to administration of zymosan and statistically analyzed by two-way ANOVA, followed by Tukey’s multiple comparisons test. ^# is for p < 0.0.5, ^## is for p < 0.01, and ^### is for p < 0.001 versus V. N = 7/group.

Tolerance and Withdrawal.

The studies above demonstrate that MJ2 and MJ5 are potent, long-lasting analgesic drugs in multiple pain types. However, these tests do not demonstrate a decrease in side effects, as hypothesized from our rationale to develop opioid pan agonists. We thus selected MJ2 for further side effect testing; we did so because MJ2 demonstrated the best balanced profile of opioid pan agonism taking into account the in vitro studies from Figures 2–3, especially since MJ2 demonstrated a higher KOR efficacy than MJ5 in Figure 3.

To begin, we first established a systemic s.c. dose that we could use for all testing. In acute tail flick studies, we found that 3.2 mg/kg s.c. MJ2 produced a mid-efficacy response suitable for repeated dosing, neither too low to produce any response nor too high to overwhelm the system and produce confounding results (Figure 8A). We then injected MJ2 at this dose twice a day for 3 days, with daily tail flick measurements. We found that MJ2 produced rapid tolerance beginning on day 2 that was down to vehicle-treated BL levels by day 3 (Figure 8B). This result suggests that MJ2 does not have benefits in tolerance, and this tolerance development is on par with morphine and similar opioids.²⁶ However, we also tested naloxone-precipitated withdrawal of MJ2, measuring both dependence and withdrawal. We found that MJ2 produced no jumping behavior, the most robust marker for withdrawal in mice, although it did produce enhanced urine/feces output (Figure 8C). In contrast, a morphine comparison control (also at 3.2 mg/kg) produced measurable jumping behavior with comparable urine/feces output. These results suggest that MJ2 has decreased dependence and withdrawal liability in comparison to morphine, supporting our pan-agonist hypothesis, although it still does display antinociceptive tolerance.

Figure 8.

Tolerance and withdrawal liability of MJ2. CD-1 mice were used for all experiments, N = 5–10/group. Data represented as the mean ± SEM. (A) Mice were tested for BL tail flick response, 52 °C, 10 s cutoff, prior to MJ2 (3.2 mg/kg) or vehicle injection s.c. and a tail flick time course. At this dose, MJ2 produced moderate- to high-efficacy antinociception, suitable for repeated dosing. (B) MJ2 (3.2 mg/kg, s.c.) or vehicle was injected twice daily with a tail flick measurement taken 30 min after the morning injection (day 1 = acute injection response, no repeated dosing). MJ2 produced progressive tolerance, down to vehicle levels by day 3. (C) Vehicle, morphine (3.2 mg/kg, s.c.) or MJ2 (3.2 mg/kg, s.c.) dosed as for tolerance studies above. On day 4, another injection was given, and 1 h later, naloxone (30 mg/kg, i.p.) was given to precipitate withdrawal. Jumps were recorded for 20 min, with urine/feces weighed at the end of the observation period. * = p < 0.05 vs vehicle group by one-way ANOVA with Dunnett’s post hoc test.

Overall, this structure–activity study proposed to probe the influence of inserting a Dmt^1,1′ residue into novel linear and cyclic analogues of the opioid peptides biphalin and AM94. The easy and straightforward incorporation of 2,5-dimethyl tyrosine in place of the native Tyr was achieved following a well-established solution-phase peptide synthesis protocol, leading to the novel compounds MJ1–5 with good overall yields and excellent purity (>95%). These compounds were investigated in competition binding assays on the three opioid receptors in vitro, showing the following trend in affinity: DOR > MOR > KOR. Both MJ1 and MJ2 retain equal affinity at MOR, while MJ3 and MJ4 lose that affinity; however, MJ3 is the most DOR-selective (16-fold vs MOR and about 230-fold vs DOR). Peptide MJ5 is an excellent binder at the three opioid receptors with 10–20 nM affinity. All compounds lose affinity at KOR with the exception of MJ2 and MJ5, which makes them truly non-selective ligands. This profile builds off our earlier generations of compounds such as the MACE series which show similar MOR and DOR affinity but lose affinity at KOR.¹⁹ We have thus generated non-selective opioid ligands in MJ2 and MJ5 which may result in improved analgesia with reduced side effects.

All compounds also showed high-potency and high-efficacy agonism at MOR with MJ1 and 2 having subnanomolar potency with good efficacy for all (78–90%). All compounds also showed high-potency partial agonism at DOR with the exception of MJ3, which showed no activity. Considering the high DOR binding affinity of MJ3, this may mean that the compound is a competitive antagonist. Again, this profile is similar to earlier generations of compounds such as the MACE series.¹⁹

Similar to the binding affinity, where these compounds differed from previous generations, was their KOR activity. MJ3 and MJ5 had weaker potency KOR partial agonism, similar to past compounds. MJ1, MJ2, and MJ4 had more potent agonist activity, which differentiated on efficacy; MJ4 was very weak (~30%), while MJ1 was stronger (~50%) and MJ2 had the highest KOR efficacy of the series (~75%). This makes these compounds non-selective opioid ligands as proposed, with MJ2 being a particularly strong candidate due to its higher KOR efficacy.

Since MJ1 is the analogue of biphalin containing Dmt^1,1′ in place of Tyr^1,1′, it is feasible to compare its activity to that of the parent compound; they show very similar K_I values against MOR and DOR, while the agonist activity of MJ1 is stronger than that of the parent compound on MOR.¹⁸ The fluorinated analogue of biphalin, AM94, exhibits extraordinary affinity for MOR and DOR (K_Iμ: 0.11 nM and K_Iδ: 0.09 nM) and extremely potent agonist activity in the GTPγS binding assay (EC₅₀: 1 nM, E_max: 77% for MOR and EC₅₀: 0.8 nM, E_max: 94% for DOR), whereas our novel linear peptide MJ2 is able to bind all the three opioid receptors with high affinity, but it is capable of stimulating only MOR with an EC₅₀ value higher than that of AM94, while the potency at DOR is very close to that of the parent compound.¹⁸ The cyclic peptide MJ3 incorporated the Dmt^1,1′ residue in place of Tyr present in the first cyclic analogue of biphalin described by Mollica et al.²⁵ This compound shows lower binding affinity for MOR/DOR compared to the parent compound (K_Iμ: 0.6 nM, K_Iδ: 0.87 nM); however, it exhibits a stronger efficacy on MOR (EC₅₀: 7.6 nM, E_max: 78% for MJ3 against EC₅₀: 0.2 nM, E_max: 47% for the parent compound). Interestingly, we observed a modest binding affinity and potency of this compound for KOR. Peptide MJ4 incorporates the same o-xylene bridge of the previously described cyclic compound MACE3,¹⁹ retaining also the p-F-Phe^4,4′ and piperazine linker. The novel compound is able to bind DOR with higher affinity than the parent peptide but very close potency and efficacy at both MOR and DOR, with poor efficacy for KOR compared with the GTPγS stimulation assay of MACE3. The cyclic peptide MJ5 preserves the same o-xylene bridge, Phe^4,4′ and hydrazine linker found in peptide 6a described by us;²² the novel peptide shows lower binding affinity for MOR and DOR with the appearance of KOR affinity; however, we observed a subnanomolar potency at DOR and a strong agonist efficacy at MOR. It is worth noting that the newly designed peptide MJ5 is an effective trivalent ligand showing K_I values close to each other, well balanced MOR/DOR/KOR efficacy, and strong potency at MOR and DOR instead of MACE4, which is a bivalent MOR/DOR agonist.

These in vitro observations also explained some of the results of our in vivo testing. Out of the whole series, MJ2 and MJ5 had the most consistently high potency, high efficacy, and long-lasting analgesic effect. This was across both i.c.v. and local administration and across several pain types, including nociceptive and inflammatory pain. This may be due to their balanced binding profile, as the most non-selective pan agonists of the series. In addition, other factors may have contributed to this observation, such as good (or poor) pharmacokinetics. This is particularly likely for MJ4, which had poor performance in vivo across the board, despite strong in vitro MOR potency and efficacy. Notably, although peptidic in structure, several of these compounds produced high-efficacy, long-lasting antinociception after systemic administration (s.c.). This also strongly implies good pharmacokinetics and metabolic stability and likely the ability to cross the blood–brain barrier. Last, MJ2 showed tolerance but less dependence and withdrawal liability, which supports the hypothesis that pan agonists have a beneficial therapeutic profile.

In this work, we describe the creation of the MJ series of peptides based on the biphalin and AM94 scaffolds. By varying the residues and cyclic linkages, we were able to design a series with improved KOR affinity and activity versus earlier generations, moving closer to the creation of a truly balanced trivalent pan agonist. All compounds showed at least some antinociceptive activity in mice, with several showing high-potency, high-efficacy, long-lasting analgesia after systemic administration by several routes. One of our analogues, MJ2, also showed decreased withdrawal liability, in line with the pan agonist hypothesis.

EXPERIMENTAL SECTION

Methods and General Procedures.

The designed products were prepared via solution-phase peptide synthesis, following the procedure previously reported by us.¹⁹ The intermediate TFA salts were used for subsequent reactions without further purification. The reference compounds biphalin and BS-S were prepared for biological purposes.^18,25 The final peptides as TFA salts were purified by reversed-phase high-performance liquid chromatography (RP-HPLC): Waters XBridgeTM Prep BEH130 C18, 5.0 lm, 250 mm 9 10 mm column; HPLC solvent A, 0.1% TFA in water; solvent B, acetonitrile; gradient: 5–95% B in A over 32 min; flow rate 5.0 mL/min. N^α-Boc-protected and Fmoc-protected intermediates were characterized by low-resolution mass spectroscopy (LRMS) and ¹H NMR analysis, following the procedures previously described by us.¹⁹ LRMS spectra were recorded using an LCQ Finnigan-Mat mass spectrometer (San Jose, CA) with an ESI-spray source and ion trap analyzer, the capillary temperature at 200 °C, and the spray voltage at 4.00 kV. Nitrogen (N₂) and helium were used as both the sheath gas and the auxiliary gas. ¹H NMR spectra were recorded at 25 °C on a 300 MHz Varian Oxford spectrometer, with DMSO-d₆ as a solvent and chemical shifts in parts per million (δ) downfield from TMS. The purity and the retention time (t_R) of the final TFA salts were determined by analytical RP-HPLC (C18-bonded 4.6 × 150 mm), flow rate 1 mL^.min⁻¹, eluent: H₂O/acetonitrile 0.1% TFA from 10% acetonitrile to 90% acetonitrile in 30 min; the purity was found to be ≥ 95% (UV detection at 214 nm). The final structures were confirmed by ¹H NMR analysis (Varian Mercury 300 MHz) and mass spectrometry ESI–HRMS (see the Supporting Information). Proton signals were assigned performing TOCSY-2D and COSY-2D experiments. The cyclic and linear peptides were obtained as TFA salts in good overall yield and ≥95% purity following RP-HPLC purification (see the Supporting Information).

2 TFA (Dmt-d-Ala-Gly-Phe)₂-hydrazine (MJ1).

91% overall yield. t_R (HPLC) = 15.30 min ¹H NMR (300 MHz, DMSO-d₆): δ 10.20 (d, 2H, NH-NH), 9.08 (s, 2H, Dmt OH), 8.51 (br, 2H, Phe NH), 8.35 (br, 6H, Dmt NH₃⁺), 8.11 (t, 2H, Gly NH), 8.06 (br, 2H, D-Ala NH), 7.36 (d, 2H, Dmt aromatics), 7.27 (d, 2H, Dmt aromatics), 7.21–6.55 (m, 10H, Phe aromatics), 4.85 (m, 2H, Phe α-CH), 4.72 (m, 2H, Dmt α-CH), 4.55 (m, 2H, D-Ala α-CH), 3.76–3.54 (dd, 4H, Gly α-CH), 3.44–3.19 (m, 8H, Dmt, Phe β-CH₂), 2.34 (s, 12H, Dmt CH₃*4), 1.48 (d, 6H, D-Ala CH₃*2). HRMS calcd for C₅₀H₆₄N₁₀O₁₀ (m/z), 964.4807; found, 965.4822 [M + H]⁺.

2 TFA (Dmt-d-Ala-Gly-P-F-Phe)₂-piperazine (MJ2).

84% overall yield. t_R (HPLC) = 15.98 min ¹H NMR (300 MHz, DMSO-d₆): δ 9.08 (s, 2H, Dmt OH), 8.53 (br, 2H, p-F-Phe NH), 8.35 (br, 6H, Dmt NH₃⁺), 8.17 (t, 2H, Gly NH), 8.06 (br, 2H, D-Ala NH), 7.36 (d, 2H, Dmt aromatics), 7.27 (d,2H, Dmt aromatics), 7.27–6.76 (m, 8H, p-F-Phe aromatics), 4.76 (m, 2H, p-F-Phe α-CH), 4.72 (m, 2H, Dmt α-CH), 4.56 (m, 2H, D-Ala α-CH), 3.76–3.54 (dd, 4H, Gly α-CH), 3.48–3.19 (m, 16H, Dmt, p-F-Phe β-CH₂, (–CH₂–CH₂–)₂), 2.34 (s, 12H, Dmt CH₃*4), 1.48 (d, 6H, D-Ala CH₃*2). HRMS calcd for C₅₄H₆₈F₂N₁₀O₁₀, (m/z), 1054.5088; found, 1055.5105 [M + H]⁺.

2 TFA c(Dmt-d-Cys-Gly-Phe)₂-hydrazine (MJ3).

61% overall yield. t_R (HPLC) = 15.95 min ¹H NMR (300 MHz, DMSO-d₆): δ 10.16 (d, 2H, NH-NH), 9.08 (s, 2H, Dmt OH), 8.90 (br, 2H, D-Cys NH), 8.51 (br, 2H, Phe NH), 8.35 (br, 6H, Dmt NH₃⁺), 8.11 (t, 2H, Gly NH), 7.36 (d, 2H, Dmt aromatics), 7.27 (d, 2H, Dmt aromatics), 7.21–6.76 (m, 8H, Phe aromatics), 4.85 (m, 2H, Phe α-CH), 4.72 (m, 2H, Dmt α-CH), 3.89 (m, 2H, D-Cys α-CH), 3.76–3.54 (dd, 4H, Gly α-CH), 2.88–2.45 (m, 8H, Phe, Dmt β-CH₂ + 4H, D-Cys β-CH₂), 2.34 (s, 12H, Dmt CH₃*4). HRMS calcd for C₅₀H₆₂N₁₀O₁₀S₂ (m/z), 1026.4092; found, 1027.4177 [M + H]⁺.

2 TFA c(Dmt-d-Cys-Gly-P-F-Phe)₂-piperazine o-Xylene (MJ4).

75% overall yield. t_R (HPLC) = 16.90 min ¹H NMR (300 MHz, DMSO-d₆): δ 9.08 (s, 2H, Dmt OH), 8.62 (br, 2H, D-Cys NH), 8.53 (d, 2H, p-F-Phe NH), 8.35 (br, 6H, Dmt NH₃⁺), 8.21 (t, 2H, Gly NH), 7.38–6.66 (m, 12H, aromatics), 6.73 (d, 4H, Dmt aromatics), 4.95 (m, 2H, p-F-Phe α-CH), 4.76 (m, 2H, Dmt α-CH), 4.16 (m, 2H, D-Cys α-CH), 3.81–3.61 (m, 4H, Gly α-CH), 3.47–3.15 (m, 8H, –CH₂–CH₂–), 3.34 (m, 4H, Dmt β-CH₂), 3.12–2.76 (m, 4H, p-F-Phe β-CH₂ + 4H, D-Cys β-CH₂), 2.47 (s, 4H, –CH₂–), 2.34 (s, 12H, Dmt CH₃*4). HRMS calcd for C₆₂H₇₄F₂N₁₀O₁₀S₂, (m/z), 1220.4999; found, 1221.5118 [M + H]⁺.

2 TFA c(Dmt-d-Cys-Gly-Phe)₂-hydrazine o-Xylene (MJ5).

71% overall yield. t_R (HPLC) = 16.38 min ¹H NMR (300 MHz, DMSO-d₆): δ 10.16 (d, 2H, NH-NH), 9.08 (s, 2H, Dmt OH), 8.76 (br, 2H, D-Cys NH), 8.28 (d, 2H, Phe NH), 8.35 (br, 6H, Dmt NH₃⁺), 8.06 (t, 2H, Gly NH), 7.35–7.01 (m, 14H, aromatics), 7.36 (d, 4H, Dmt aromatics), 4.81–4.66 (m, 2H, Phe α-CH + m, 2H, Dmt α-CH), 4.11 (m, 2H, D-Cys α-CH), 3.43–3.26 (dd, 4H, Gly α-CH), 3.45–2.91 (m, 4H, Dmt β-CH₂), 2.82–2.41 (m, 4H, Phe β-CH₂ + 4H, D-Cys β-CH₂), 2.47 (s, 4H, –CH₂–), 2.34 (s, 12H, Dmt CH₃*4). HRMS calcd. For C₅₈H₇₀N₁₀O₁₀S₂ (m/z), 1130.4718; found, 566.2435 [M/2]⁺⁺.

ABBREVIATIONS

MOR

μ-opioid receptor

DOR

δ-opioid receptor

KOR

κ-opioid receptor

SAR

structure–activity relationships

DAMGO

[DAla(2), N-Me-Phe-(4), Gly-ol(5)] enkephalin

DPDPE

D-Pen2,D-Pen5-enkephalin

TFA

trifluoroacetic acid

i.c.v.

intracerebroventricular

i.v.

intravenous

s.c.

subcutaneous

DMSO

dimethylsulfoxide

MPE

maximum possible effect

RP-HPLC

reversed-phase high-performance liquid chromatography

FT

formalin test

HP

hot plate

BL

baseline

i.p.

intraperitoneal

LRMS

low-resolution mass spectroscopy

ESI-HRMS

high-resolution electrospray ionization mass spectrometry

TMS

tetramethylsilane

NMR

nuclear magnetic resonance