Design, Synthesis, and In Vitro Characterization of Proteolytically-Stable Opioid-Neurotensin Hybrid Peptidomimetics

Abstract

Linking an opioid to a nonopioid pharmacophore represents a promising approach for reducing opioid-induced side effects during pain management. Herein, we describe the optimization of the previously reported opioid-neurotensin hybrids (OPNT-hybrids), SBL-OPNT-05 & -10, containing the μ-/δ-opioid agonist H-Dmt-d-Arg-Aba-β-Ala-NH₂ and NT(8–13) analogs optimized for NTS2 affinity. In the present work, the constrained dipeptide Aba-β-Ala was modified to investigate the optimal linker length between the two pharmacophores, as well as the effect of expanding the aromatic moiety within constrained dipeptide analogs, via the inclusion of a naphthyl moiety. Additionally, the N-terminal Arg residue of the NT(8–13) pharmacophore was substituted with β³hArg. For all analogs, affinity was determined at the MOP, DOP, NTS1, and NTS2 receptors. Several of the hybrid ligands showed a subnanomolar affinity for MOP, improved binding for DOP compared to SBL-OPNT-05 & -10, as well as an excellent NTS2-affinity with high selectivity over NTS1. Subsequently, the G_αi1 and β-arrestin-2 pathways were evaluated for all hybrids, along with their stability in rat plasma. Upon MOP activation, SBL-OPNT-13 and -18 were the least effective at recruiting β-arrestin-2 (E_max = 17 and 12%, respectively), while both compounds were also found to be partial agonists at the G_αi1 pathway, despite improved potency compared to DAMGO. Importantly, these analogs also showed a half-life in rat plasma in excess of 48 h, making them valuable tools for future in vivo investigations.

Chronic pain, which affects approximately 20% of the world’s adult population, stands among the most important societal burdens. Not only does this condition cause major health problems, but it also has a considerable impact on patients’ quality of life and societal engagement. Yet there is still no satisfactory and adequate way of treating pain.¹ Currently, opioids such as morphine, oxycodone, and fentanyl are used to treat chronic pain but unfortunately, their strong analgesic effects are accompanied by serious side effects, such as nausea, sedation, constipation, respiratory depression, physical dependence, and tolerance.² Moreover, the occurrence of opioid use disorders due to abusive prescription practices, falling prices, and increased accessibility of synthetic opioids has significantly contributed to the persistent “opioid crisis”.³ To overcome these side effects, three main strategies have been developed in recent decades: (i) the development of G protein-biased μ-opioid receptor (MOP) agonists,^4,5 (ii) increasing affinity and activity toward the δ-opioid receptor (DOP), and (iii) the design of bifunctional ligands, also known as hybrids or designed multiple ligands.^2,6,7 In recent years, G protein-biased μ-opioid agonists have been suggested to prevent opioid-induced adverse effects, but recent data have revealed that these adverse effects are not only mediated by β-arrestin recruitment but also by G protein-dependent signaling pathways.⁸⁻¹⁴ Even though the impact of G protein-biased agonism is still debated,⁹⁻¹² partial agonism at the μ-opioid receptor (i.e., low intrinsic efficacy agonists) has been suggested to be favorable for the development of safer analgesics with more acceptable side effect profiles. Importantly, the deconvolution of the biological responses generated by G protein-coupled receptor (GPCR) signaling pathways remains a key challenge in linking in vitro to in vivo responses.¹⁵⁻¹⁷ A second strategy involves the development of MOP/DOP bifunctional ligands. While maintaining MOP agonism, simultaneous targeting of DOP can lead to beneficial effects.² DOP is known to also produce analgesic effects and it is involved in the regulation of both antidepressant and anxiolytic-like effects, and mood disorders.^2,18 Anxiety and mood disorders can be associated with chronic pain, making DOP a desirable target for long-term pain treatment.^2,18 Highly selective DOP agonists may therefore present a favorable benefit–risk profile: antidepressant and anxiolytic-like effects, antihyperalgesia, less tolerance, and less respiratory depression,^2,19 the latter three being the main side effects induced by opioids. Taken together, this indicates that high-affinity and biased signaling at DOP represents a promising approach. The last strategy consists of developing bifunctional ligands, also called hybrids, multitarget or designed multiple ligands, in which an opioid and a nonopioid pharmacophore can, for instance, be covalently linked. They offer many advantages over drug cocktails (i.e., administration of multiple drugs); their pharmacokinetic and pharmacodynamic properties are more predictable and the risks of drug interactions are lower. Additionally, while maintaining agonism at one receptor, targeting a second receptor can lead to attenuated side effects.^2,20 One of the composite pharmacophores of the hybrid ligands would be responsible for interaction with MOP as well as with other opioid receptors,²⁰ while the nonopioid pharmacophore could then regulate opioid-induced side effects or show opioid-independent actions.²¹⁻²⁴ Among nonopioid receptors, neurotensin (NT) and its receptors (NTS1 and NTS2) produce antinociceptive responses that are independent of the opioidergic system.²⁵ Furthermore, the coadministration of opioid and NTS agonists has been reported to produce an additive and even synergistic action.^26,27 Moreover, in a preclinical neuropathic pain model, NT analogs have been shown to be capable of reversing nociception, whereas morphine struggles to treat this type of pain.²⁸ The fusion of an NT moiety to an opioid ligand could thus offer a way to treat a wider spectrum of pain conditions. In terms of structure, the naturally occurring NT is a tridecameric neuropeptide with the sequence pGlu-Leu-Tyr-Glu-Asn-Lys-Pro-Arg-Arg-Pro-Tyr-Ile-Leu-OH. It was discovered in 1973 by Carraway and Leeman,²⁹ and acts through its three known receptors: NTS1, NTS2, and NTS3. The first two receptors belong to the GPCR superfamily, while the latter one, also known as sortilin, is a single transmembrane domain receptor.³⁰ Ever since the discovery of the minimal active-sequence of NT (i.e., H-Arg-Arg-Pro-Tyr-Ile-Leu-OH; NT(8–13)),^31,32 many analogs have been developed.³³⁻⁴⁰ Importantly, while the antinociceptive activity is primarily related to NTS1/NTS2 binding and activation, NTS1 also promotes hypotension and hypothermia.^35,41 This has served as an incentive to develop NTS2-selective ligands in the context of pain management.³⁹ Practically, the biological activity of NT analogs is exerted through recognition of its C-terminus, while the N-terminus of opioid peptide ligands is key for activation of the MOP and DOP receptors. Consequently, opioid and NT pharmacophores could be readily assembled by means of solid-phase peptide synthesis (SPPS). The first opioid-neurotensin (OPNT) hybrid described, PK20 (H-Dmt¹-d-Lys²-Phe³-Phe⁴-Lys⁵-Lys⁶-Pro⁷-Phe⁸-Tle⁹-Leu¹⁰-OH, where Dmt = 2′,6′-dimethyl-tyrosine), was disclosed in 2010, displaying strong antinociceptive effects after central and peripheral administration.⁴² In an effort to improve this particular ligand, our groups reported on a follow-up series of OPNT hybrids,²³ taking PK20 as the lead peptide sequence. This work resulted in two hybrids, each with high affinity for all three opioid receptors and excellent selectivity for NTS2 over NTS1 (Figure 1). Of these two, SBL-OPNT-05 also demonstrated a much longer in vivo analgesic response (up to 8 h) than its opioid parent compound (KGOP01—H-Dmt-d-Arg-Aba-β-Ala-NH₂, Aba = tetrahydro-4-amino-2-benzapinone) and PK20.²³ It should be pointed out that only the NT-pharmacophore was optimized in the previously disclosed work, with the incorporation of Dmt⁸ and β³hArg⁵ in SBL-OPNT-05 and Tic(6-OH)⁸ in SBL-OPNT-10, beside the introduction of tert-Leu⁹ (Tle) in both hybrids.

Figure 1.

Structures of the two lead hybrids SBL-OPNT-05 and SBL-OPNT-10 (including in vitro pharmacological data) with the opioid pharmacophore framed in orange and the NT pharmacophore framed in blue.

Consequently, we present herein a second generation of OPNT hybrids derived from SBL-OPNT-05 and -10 (Figure 1). Further structure–activity relationships were obtained by varying the conformationally constrained dipeptide (in positions 3 and 4) such that the steric hindrance, as well as the distance between the two pharmacophores, were altered. The required noncanonical amino acids were synthesized and subsequently incorporated into the desired peptide sequences. The ligands were then evaluated in vitro by establishing their affinities for MOP, DOP, NTS1, and NTS2, using competitive radioligand binding assays. Thereafter, the potency and efficacy of both MOP and DOP were assessed by BRET assays to monitor G_αi1 activation and β-arrestin-2 recruitment. Finally, their proteolytic stability was investigated in rat plasma.

Results and Discussion

Design and Synthesis

As described above, SBL-OPNT-05 and -10 displayed nanomolar binding to MOP and excellent NTS2 selectivity (1333- and 2739-fold, respectively) in addition to nanomolar MOP potency toward the G protein pathway (EC₅₀ = 28.7 and 4.56 nM), and reduced β-arrestin-2 efficacy (E_max = 68 and 67%, respectively). Importantly, SBL-OPNT-05 showed a half-life of almost 8 h.²³ We hypothesized, therefore, that changes in opioid pharmacophore could lead to additional improvements to the molecules’ signaling and stability profiles. In order to achieve this, we synthesized a novel series of peptidomimetics, in which the constrained dipeptide at positions 3 and 4 has been modified. Hence, the Aba-β-Ala segment was modified to include not only naphthyl analogs, but also spacer variations (from a shorter linking amino acid (Gly) to γ-aminobutyric acid (GABA) as the extended spacer). This variation aimed to modulate both opioid receptor affinity and in vitro activity, as well as altering the molecules’ functional selectivity (i.e., G protein signaling versus β-arrestin recruitment). In position 5, both Arg and β³hArg were inserted as commercially available building blocks. All derived peptides contain Dmt at the N-terminal position, on account of its ability to improve the affinity and activity of opioid peptides.⁴³ Although this noncanonical residue is commercially available, it was synthesized due to its high commercial cost (Scheme 1). However, the substitution of Dmt in position 8 by Tic(6-OH) led to a better selectivity toward NTS2. In addition, building blocks containing an Aba or a tetrahydro-4-amino-2-naphthazepinone (Ana) unit were also synthesized. These amino acids correspond to conformationally constrained variants of Phe and 1-naphthylalanine (1Nal), respectively. The synthesis of the Aba moiety was previously developed by our lab^44,45 and subsequently extended to the preparation of the 1Ana moiety.⁴⁶ In order to influence signaling and to modulate steric constraints between pharmacophores, six different dipeptides were synthesized (i.e., Aba-Gly, Aba-β-Ala, Aba-GABA, 1Ana-Gly, 1Ana-β-Ala and 1Ana-GABA). Depending on the amino acid in position 4, slightly different strategies are used (shown in Schemes 2 and 3). As the last amino acid to be synthesized, Fmoc-Tic(6-OH)-OH was prepared according to the synthetic protocol described in Gonzalez et al.²³

Scheme 1. Synthesis of Boc-Dmt-OH.

(a) SOCl₂ (1.1 equiv), MeOH (anhydrous, 1.0 M), reflux, o.n., quant.; (b) TBDMSCl (1.2 equiv), imidazole (1.2 equiv), CH₂Cl₂ (anhydrous, 0.2 M), r.t., o.n., 97%; (c) picolinic acid (1.2 equiv), TBTU (1.2 equiv), DIPEA (2.5 equiv), CH₂Cl₂ (anhydrous, 0.2 M), r.t., o.n., 73%; (d) MeI (5.0 equiv), K₂CO₃ (3.0 equiv), Pd(OAc)₂ (5 mol %), toluene (0.25 M), 120 °C, 48 h; 69%; (e) HCl_conc (17.0 equiv), H₂O (0.7 M), 0 °C, 90 °C, 24 h; (f) Boc₂O (1.1 equiv), Et₃N (1.1 equiv), THF, r.t., o.n., 54% (over two steps).

Scheme 2. Synthesis of Fmoc-Aba-Gly-OH and Fmoc-1Ana-Gly-OH.

(a) Methyl 2-(succinimidooxy)carbonyl-benzoate (1.1 equiv), Na₂CO₃ (3.0 equiv), MeCN/H₂O (3:2 v/v, 0.05 M), r.t., o.n., 90 and 96%; (b) HCl H-Gly-OEt (1.1 equiv), TBTU (1.1 equiv), Et₃N (3.0 equiv), CH₂Cl₂ (anhydrous, 0.2 M), r.t., 1 h 30 min, 94 & 65%; (c) 1 N HCl/acetone (1:1 v/v, 0.05 M), 90 °C, 4h30, quant. and 92%; (d) paraformaldehyde (15.0 equiv), pTosOH (0.1 equiv), toluene (0.05 M), reflux, 1 h 30 min, 93 and 73%; (e) triflic acid (5.0 equiv), CH₂Cl₂ (anhydrous, 0.1 M), r.t., 1 h, argon, quant. and quant.; (f) hydrazine (6.6 equiv), EtOH (0.1 M), reflux, 30 min; (g) FmocOSu (1.1 equiv), Na₂CO₃ (1.2 equiv), H₂O/acetone (1:1 v/v, 0.1 M), r.t., o.n., 64 and 48% (over two steps).

Scheme 3. Synthesis of Fmoc-Aba-β-Ala-OH (n = 1), Fmoc-1Ana-β-Ala-OH (n = 1), Fmoc-Aba-GABA–OH (n = 2) & Fmoc-1Ana-GABA–OH (n = 2).

(a) Methyl 2-(succinimidooxy)carbonyl-benzoate (1.1 equiv), Na₂CO₃ (3.0 equiv), MeCN/H₂O (3:2 v/v, 0.05 M), r.t., o.n., 90 and 96%; (b) HCl H-β-Ala-OEt or HCl H-GABA-OEt (1.1 equiv), TBTU (1.1 equiv), Et₃N (3.0 equiv), CH₂Cl₂ (anhydrous, 0.2 M), r.t., 1 h 30 min, quant. & 74% (for Phe), 65 and 94% (for 1-Nal); (c) 1,3,5-trioxane (6.6 equiv), P₂O₅ (22.2 equiv), H₃PO₄ (85% aq., 10.7 equiv), benzene/AcOH (3:2 v/v, 0.025 M), 115 °C, 1 h 30 min, 63 and 79% (for Phe), 82 and 65% (for 1-Nal); (d) 1 N HCl/acetone (1:1 v/v, 0.05 M), 90 °C, o.n., 87 & 91% (for Phe), quant. and 75% (for 1-Nal); (e) hydrazine (6.6 equiv), EtOH (0.1 M), reflux, 30 min; (f) FmocOSu (1.1 equiv), Na₂CO₃ (1.2 equiv), H₂O/acetone (1:1 v/v, 0.1 M), r.t., o.n., 37 and 59% (for Phe), 54 and 76% (for 1-Nal) (over two steps).

The synthesis of Dmt started by triple protection of Tyr (Scheme 1). First, esterification was performed, which was followed by silylation of the phenol moiety and, finally, protection of the amine using picolinic acid. This picolinamide serves as a directing group for the following Pd-catalyzed aryl-dimethylation.^47,48 A 5-fold excess of methyl iodide was required to favor the formation of the desired 2,6-dimethylated compound. Subsequently, the three protecting groups were removed in a single step under acidic conditions. Finally, the amine was protected (by means of a Boc group given its use as the N-terminal residue in this subset of analogs), yielding an SPPS-compatible building block.

The synthesis of conformationally constrained dipeptides bearing a Gly in the second position of the dipeptide moiety is shown in Scheme 2. In this case, the synthesis took Phe or 1Nal as a starting point, where the amine was protected with a phthaloyl group. Next, TBTU-mediated peptide-bond formation with H-Gly-OEt yielded the desired dipeptides, followed by hydrolysis of its ester. Subsequently, oxazolidinone formation was achieved using paraformaldehyde in the presence of a catalytic amount of acid under Dean–Stark conditions. The penultimate step in this synthesis is a cyclo–isomerization reaction; in the presence of triflic acid and through formation of an N-acyliminium ion intermediate, a seven-membered 4-amino-benzazepinone (Aba) or 4-amino-naphthazepinone (Ana) was formed, after which the phthaloyl group could be removed using hydrazine and the amine was reprotected with a Fmoc group to obtain the desired SPPS-compatible building blocks.

The synthesis of constrained dipeptides bearing β-Ala or GABA in the second position could be achieved with the cyclization in a single step (Scheme 3), with trioxane serving as the formaldehyde source. The resulting ester was then hydrolyzed and the phthaloyl group was cleaved using the same conditions described above, followed by final Fmoc protection.

Once these constrained amino acids/dipeptides were in hand, they were introduced into the envisaged hybrid peptides. All hybrids were synthesized by Fmoc/tBu SPPS methodology using preloaded Fmoc-Leu-Wang resin as solid support. However, HBTU/DIPEA was used as the coupling mixture for the initial coupling of Tle, but subsequent coupling steps were performed using DIC/Oxyma Pure in order to avoid phenol acylation since Dmt and Tic(6-OH) contain a free hydroxyl group. All peptides were cleaved from the solid support using a cleavage mixture of TFA/TIS/H₂O 95:2.5:2.5 and purified using preparative HPLC, yielding the desired compounds with excellent purity (>95%) and yields ranging from 5 to 37% (see Table S1 in Supporting Information). The investigated hybrids are shown in Table 1.

Table 1. OPNT Hybrid Peptides Investigated in This Study^a.

^aChemical modifications with respect to the lead sequences are indicated in bold. Reference compounds are highlighted by a grey background.

In Vitro Binding Affinity and Selectivity

The binding affinity of the newly synthesized hybrid peptides was determined at human μ- & δ-opioid receptors (MOP & DOP) and NT-1 and 2 receptors (NTS1 & NTS2) using competitive radioligand binding assays (Figures S1 and S2). For opioid receptors, membrane extracts from HEK293 cells stably expressing hMOP or hDOP were used. For NTS receptors, CHO-K1 cells stably expressing hNTS1 or 1321N1 cells stably expressing hNTS2 were selected. The radioligand used to determine binding affinity was ¹²⁵I-[Tyr³]-NT for both NT receptors and [¹²⁵I]-DAMGO and [¹²⁵I]-deltorphin I for MOP and DOP, respectively. All binding affinity data are expressed as K_i values and shown in Table 2.

Table 2. Binding Affinities of OPNT Hybrids to Human Opioid and NT Receptors, in Addition to Selectivity toward NTS2 over NTS1^e.

^aDetermined in competitive radioligand binding assays using HEK293 cells stably transfected with hMOP or hDOP.

^bDetermined in competitive radioligand binding assays using CHO-K1 cells stably transfected with hNTS1.

^cDetermined in competitive radioligand binding assays using 1321N1 cells stably transfected with hNTS2. K_i values are reported as the means ± SEM of three independent experiments performed in triplicate.

^dFrom Gonzalez et al. using CHO cells for MOP, DOP, and NTS1, and 1321N1 astrocytoma cells for NTS2.²³

^eReference sequences are highlighted with a gray background.

When comparing the two lead peptides SBL-OPNT-05 and SBL-OPNT-10, both show high selectivity for NTS2 over the NTS1. However, SBL-OPNT-05 has excellent (i.e., low picomolar) affinity for NTS2, while it also binds to NTS1 in the low nanomolar range. Given that NTS1 is the receptor linked to most of the physiological effects associated with NT administration, including hypotension and hypothermia, the nanomolar NTS1 affinity is inadequate for pain management. By switching from Dmt⁸ to Tic(6-OH) (SBL-OPNT-10), an even better selectivity could be observed and binding to NTS1 was strongly attenuated with a K_i value in the micromolar range. For this reason, in subsequent series, the Tic(6-OH) moiety was retained at position 8 of the hybrid structures.

As an initial derivation of SBL-OPNT-10, the conformationally constrained Aba residue in position 3 was substituted with the sterically more demanding 1Ana system (SBL-OPNT-15). This change was only tolerated by MOP (K_i = 0.25 nM), with the affinity of DOP and NTS2 lowered by 2.5- or 3.5-fold respectively, demonstrating that this change alone was not advantageous when applying the β-Ala linking residue in position 4. However, upon introduction of a shortened linker between the two pharmacophores (i.e., Gly; SBL-OPNT-13 and -16), taking SBL-OPNT-10 as the reference analog, satisfactory affinities for μ- and δ-opioid receptors were displayed, meaning that this hybrid contraction did not impair MOP affinity, and even slightly improved DOP binding, compared with the reference sequence SBL-OPNT-10. At the same time, NTS2 binding was preserved for both analogs, and NTS1 binding was even halved in the case of SBL-OPNT-16. In the latter, the substitution of β-Ala for Gly was not the sole modification, but the aromatic ring system was also changed from an Aba to a 1Ana moiety. Comparison of SBL-OPNT-13 and -16 shows that NTS1 binding can even be beneficially lowered by insertion of the naphthoazepinone bicycle (NTS2/NTS1 selectivity is ca. 2400-fold). The same trend could be observed upon the introduction of a more extended linker (GABA; SBL-OPNT-14 & -17). On the opioid side, the insertion of an additional methylene group between the two pharmacophores did not result in any major changes; subnanomolar binding to MOP was maintained, while improved binding to DOP was observed for SBL-OPNT-14 & -17 (K_i = 12 vs 54 nM, respectively). Although, at first sight, a loss of selectivity for NTS2 was noticeable when the Aba-GABA dipeptide was inserted into SBL-OPNT-14, the K_i values for NTS2 and NTS1 remained within the same ranges, i.e., 1-digit nM for NTS2 and micromolar for NTS1. Nonetheless, the selectivity ratio improved when switching to 1Ana-GABA in SBL-OPNT-17.

Previously, the introduction of β³hLys led to high NTS2 selectivity and proteolytic stability in pure NT-analogs.³⁵ The incorporation of β³hLys⁵ into the hybrid sequences led to a drastic decrease in NTS2 selectivity. Interestingly, when β³hArg⁵ was introduced, the NTS2 selectivity was restored.²³ As a result, β³hArg was therefore introduced into a greater number of sequences. The substitution of Arg⁵(SBL-OPNT-10) by β³hArg (SBL-OPNT-12) gave rise to a 3.5-fold loss in MOP affinity, but a 2-fold increase in DOP affinity was also observed. As regards the binding to NT receptors, homologation of the Arg⁵ residue slightly reduced binding at NTS2, whereas NTS1 affinity was increased, shifting the NTS2/NTS1 selectivity in the undesired direction. A similar unfavorable decrease in selectivity was observed for SBL-OPNT-18 and -19 compared to SBL-OPNT-10, due to a significant loss in NTS2 binding and approximately 2-fold increase in NTS1 affinity. It is worth mentioning that affinity for MOP (2.2 and 3.3-fold, respectively) and DOP (1.9 and 3.6-fold, respectively) was enhanced. Both ligands showed extremely high MOP binding with K_i values of 90 and 60 pM, respectively, which is remarkable.

With the main goal of this series in mind, namely the development of dual MOP/DOP agonist ligands flanked by NTS2 selectivity, promising ligands were obtained and subnanomolar MOP binding was accompanied by slightly improved DOP affinity. In addition, the affinities of NTS2 and NTS1 were mostly unaffected by the investigated changes in the opioid pharmacophore, staying in the low nM range for NTS2 and in the μM range for NTS1. Considering these promising results, pharmacological evaluation of the ligands was pursued, and in vitro activity assessment and plasma stability tests were carried out.

In Vitro Functional Activity

After determination of their affinities, the functional activity, as well as the functional selectivity of all analogs were assessed for both MOP and DOP, including G protein signaling and β-arrestin recruitment pathways. In addition, their proteolytic stability in rat plasma was determined to predict the potential exposure of the compounds in vivo.

The activity of all compounds was established using a BRET-based assay for both G_αi1 activation and β-arrestin-2 recruitment. In the case of G_αi1 activation (Figure 2A), HEK293 cells were transfected with Flag-MOP or Flag-DOP, in addition to G_αi1-RlucII, Gβ₂, GFP10-Gγ₁. For β-arrestin-2 recruitment (Figure 2B), HEK293 were transfected with hMOP-RlucII or hDOP-RlucII, in addition to GFP10-β-arrestin-2. Of note, DAMGO and deltorphin II were used as reference ligands with full agonist activity, heading both pathways, at MOP and DOP, respectively.

Figure 2.

Concentration–response curves for hybrid ligands on the G_αi1, and β-arrestin-2 pathways following MOP activation. Schematic representation of the BRET-based assay used to assess G-protein activation showing loss of BRET² signal upon activation (A) or recruitment of β-arrestin-2 illustrating a gain in BRET² signal after receptor stimulation (B). Ligand-triggered G_αi1 engagement on MOP (C). Ligand-induced β-arrestin-2 recruitment to MOP (D). Each set represents the mean of at least three independent experiments and is expressed as mean ± SEM.

At MOP, all analogs showed greater potency for G_αi1 activation than DAMGO (Table 3), while retaining full agonism in almost all cases. Only SBL-OPNT-13 and -18 showed a slight decrease in efficacy (E_max = 87 and 85%, respectively) (Figure 2C). When measuring β-arrestin-2 recruitment by OPNT hybrids, it was noted that five analogs possessed an E_max < 80%, namely SBL-OPNT-10, -12, -13, -18 and -19 (Figure 2D). Full agonism was observed for the remaining analogs: SBL-OPNT-14, -15, -16 & -17 (E_max = 129, 112, 95 and 97%, respectively). Importantly, SBL-OPNT-13 and -18 exhibited the greatest drop in β-arrestin-2 recruitment efficacy (E_max = 17 and 12%, respectively), which is interesting in light of the putative benefits of biased signaling at MOP.^2,49,50

Table 3. Functional Activities on MOP and DOP of Hybrids^b.

^aHEK 293 cells were transfected with BRET-based biosensors to determine G_αi1 activation or β-arrestin-2 recruitment to MOP, DOP or KOP. Both signaling pathways were studied by stimulating with increasing concentrations of ligands to establish EC₅₀ and E_max. Data are reported as the mean ± SEM of three independent experiments performed in triplicate.

^bReference compounds were highlighted with grey background.

When comparing the structure of all compounds, insertion of the Aba-Gly dipeptide (SBL-OPNT-13 and -18) provides the lowest efficacy for β-arrestin-2 recruitment. Importantly, Gly seemed to be the ideal linker length between the two pharmacophores for G_αi1 activation, while no clear consensus was reached for β-arrestin-2 efficacy since the presence of 1Ana-Gly (SBL-OPNT-16) showed full agonism for both pathways. Additionally, the presence of 1Ana (SBL-OPNT-15, -16 and -17) increased the potency of G_αi1 activation.

We next assessed the potency and efficacy of G_αi1 activation and β-arrestin-2 recruitment following DOP activation by OPNT hybrids and the data were normalized to deltorphin II, a natural heptapeptide agonist of DOP (Figure 3A,B). Several members of the series (i.e., SBL-OPNT-13, -14, -16 and -17) were more potent than the reference analog SBL-OPNT-10 in activating G_αi1 at DOP, and retained full agonist activity (Table 3). In contrast, as observed with SBL-OPNT-10, all of SBL-OPNT-12, -15, -18 and -19 exhibited partial agonism toward the G_αi1 pathway. When measuring β-arrestin-2 recruitment by OPNT hybrids, SBL-OPNT-14 alone displayed a 10-fold greater potency for β-arrestin-2 recruitment than SBL-OPNT-10. Partial agonism toward the β-arrestin-2 pathway was also observed for SBL-OPNT-10, -12, -14, -15 and -18. Importantly, the favorable pharmacological activation profile of all compounds toward G_αi1 activation (except for SBL-OPNT-14) may be beneficial for future in vivo studies, as DOP is involved in antidepressant- and anxiolytic-like effects, and its relationship with chronic pain is highly desirable.^2,18

Figure 3.

Concentration–response curves for hybrid ligands on the G_αi1, and β-arrestin-2 pathways following DOP activation. DOP (A) is monitored using the BRET-based G protein dissociation assay or DOP (B) using the BRET-based β-arrestin-2 recruitment assay. Each set represents the mean of at least three independent experiments and is expressed as mean ± SEM.

Structurally, it can be concluded that SBL-OPNT-12 and -15, both bearing the β-Ala linker, exhibited the lowest potency and efficacy in recruiting β-arrestin-2 to the DOP receptor. Compared to SBL-OPNT-10 (also carrying the β-Ala linker), expanding the aromatic ring to 1Ana (SBL-OPNT-15) further reduced β-arrestin signaling (higher EC₅₀ & lowered E_max), and also worsened G_αi activation, while introduction of β³hArg (SBL-OPNT-12) appears to keep G_αi activation constant as well as reducing β-arrestin-2 signaling. The highest potency and efficacy for DOP G_αi activation was obtained for SBL-OPNT-14, which contains the Aba-GABA constraint (EC₅₀ = 81 nM, E_max = 118%). On the other hand, all compounds bearing the β³hArg (SBL-OPNT-12, -18 and -19) were partial agonists for both pathways, while compounds lacking both the β-Ala linker group and the β³hArg-moiety displayed full agonism (SBL-OPNT-13, -14, -16 and -17) at G_αi activation. SBL-OPNT-14 demonstrated the best β-arrestin-2 potency (EC₅₀ = 73 nM), which is not preferred, since we want to diminish β-arrestin-2 signaling as much as possible. Combining all signaling profiles, SBL-OPNT-13 and -18 stand out as partial agonists on G_αi activation at MOP, while establishing acceptable DOP activation with reduced β-arrestin-2 recruitment compared to deltorphin II.

Surprisingly, a high discrepancy between MOP and DOP binding or activition was observed, even though we started from a balanced MOP/DOP opioid agonist pharmacophore. Throughout this series, mainly the opioid pharmacophore and the first amino acid of the NT pharmacophore were substituted. The changes made were in most cases tolerated or beneficial for MOP binding or activation, but decreased DOP binding (SBL-OPNT-15) or activation (SBL-OPNT-12 and -15). Altogether, this suggests that DOP binding or activation is more sensitive to the substitutions made, as compared to MOP.

In contrast with the subnanomolar K_i values on MOP reported in Table 2, EC₅₀ activation values recorded in BRET-based activation assays (Table 3) were all in nanomolar range. The potency of a receptor to produce a response (measured as functional data) is driven by both affinity and efficacy at the receptor. Hence, this suggest that this series of compounds may exhibit lowered intrinsic efficacy at the opioid receptors which has recently been reported to be highly favorable for analgesics with a wider therapeutic window.⁹

Finally, the potency and efficacy of G_αi1 activation following KOP activation by the novel OPNT hybrids was assessed and the data were normalized to Dynorphin A, an endogenous full agonist opioid peptide of KOP (Figure 4). The incorporation of 1Ana (i.e., SBL-OPNT-15, -16, and -17) was highly beneficial for the activation of G_αi1 at KOP. These three hybrid ligands exhibited close to full agonist responses (E_max = 72 to 82%), in contrast to the analogues devoid of the 1Ana moiety (i.e., SBL-OPNT-10, -12, -13, -14, -18, and -19) which were all partial agonists (Table 3). The length of the linker did not impact the efficacy nor the potency to activate the G_αi1 pathway at KOP. For instance, the substitution of β-Ala at position 4 for a shorter (Gly) or longer (GABA) linker was not sufficient to yield an important shift in KOP activation. However, the incorporation of the β³hArg at the fifth position was significantly deleterious for G_αi1 pathway activation at KOP. Indeed, compounds carrying this nonproteogenic amino acid (SBL-OPNT-12, -18, and -19) exhibited a loss in potency, as well as efficacy, when compared to their Arg-bearing counterparts (SBL-OPNT-10, -13, and -14).

Figure 4.

Concentration–response curves for hybrid ligands on the G_αi1 pathways following KOP activation. Ligand-triggered G_αi1 engagement at KOP was evaluated using BRET-based biosensors. Each set represents the mean of at least three independent experiments and is expressed as mean ± SEM.

Determination of Hybrid Stability in Rat Plasma

Enzymatic degradation of NT(8–13) occurs at three primary cleavage sites, Arg⁸-Arg⁹, Pro¹⁰-Tyr¹¹, and Tyr¹¹-Ile¹² by a combination of multiple Zn metalloendopeptidases, namely the endopeptidases 24.11, 24.15, and 24.16.⁵¹ Hence, a further series of modifications was introduced into the NT(8–13) sequence to protect these metabolically susceptible spots and increase the half-life of the analogs (Figure 4A and Table 4). In 2019, Eiselt et al. described a series of pure NT analogs in which these weak spots were targeted to improve proteolytic stability. The Pro¹⁰-Tyr¹¹ and Tyr¹¹-Ile¹² bonds were protected by substituting Tyr with a cyclic m-Tyr (Tic(6-OH)) and Ile¹² was replaced by tert-leucine (Tle) to generate SBL-NT-35,³⁵ which corresponds to the NT part of SBL-OPNT-10, -13, -14, -15, -16 and -17. These modifications resulted in an increase in t_1/2 of 4.8 min over NT(8–13) (Table 4 and Figure 5A). In their report, Eiselt et al. also described the implementation of a β³hLys residue at the N-terminus, resulting in a half-life of more than 24 h.³⁵ As a result, we further protected the two basic amino acid residues by replacing Arg⁸ with a nonproteinogenic β³hArg to create SBL-NT-34, which corresponds to the NT moiety of SBL-OPNT-10, -12, -18 & -19. This NT analog displayed an improved plasma stability, offering a t_1/2 of over 15 min. Moreover, by fusing the NT moiety to each opioid tetrapeptide, the Arg⁵-Arg⁶ residues were in any case protected by the steric hindrance entailed by the adjacent opioid part at the N-terminus, resulting in a half-life of over 48 h for all hybrids (Table 4, Figure 5B).⁵² Hence, as expected, protecting all proteolytic sites afforded desired levels of plasma stability (>48 h) for all OPNT hybrid analogs.

Table 4. Plasma Stability of Each OPNT Hybrid with Respect to Its NT Moiety^a.

^aRat plasma was incubated with the OPNT hybrid entities for up to 48 h. The half-life of these compounds was calculated from the degradation curve using the one-phase exponential decay function in GraphPad Prism 9. The values represent the mean of three experiments.

Figure 5.

Plasma stability of hybrid ligands and their corresponding NT moiety compared to NT(8–13). Degradation curves were generated following incubation of the OPNT hybrids in rat plasma at 37 °C for up to 15 min (A) or 48 h (B). Each series represents the mean of three independent experiments and is expressed as mean ± SEM.

Conclusion

In conclusion, we report the synthesis of constrained aromatic amino acids/dipeptides and their incorporation into newly developed OPNT multitarget peptidomimetics. Two analogs, SBL-OPNT-13 and -18 stood out in terms of affinity, activity, and stability. They exhibited excellent binding to MOP, DOP, and NTS2 (K_i = 0.34, 37, 2.54 and 0.09, 37, 4.26 nM, respectively), in addition to excellent NTS2 selectivity over NTS1. On top of this, improved potency for G_αi1activation at MOP was observed (EC₅₀ = 7.4 and 24 nM, respectively) compared to DAMGO (EC₅₀ = 32.2 nM). Additionally, a drastic decrease in β-arrestin-2 recruitment efficacy at MOP was demonstrated (E_max = 17 and 12%, respectively), while still establishing acceptable DOP activation with reduced β-arrestin-2 recruitment compared to the reference. Importantly, their half-life in rat plasma is greater than 48 h. This provides a steady basis for future pharmacological analyses, both in vitro and in vivo, in view of the development of improved analgesics.

Experimental Section

Materials

Chemicals for building block synthesis, Fmoc-protected amino acids, and other compounds were purchased from BLD Pharm (Shanghai, China—Fmoc-β³hArg-OH, Fmoc-Tle-OH, HBTU, TIS, HCl H-GABA-OEt and FmocOSu), Fluorochem (Hadfield, UK—Fmoc-Arg(Pbf)-OH, Fmoc-d-Arg(Pbf)-OH, DIC, TBTU, TBDMSCl, triflic acid), Aldrich (St. Louis, USA—TFA, DIPEA, 4-methylpiperidine, SOCl₂, imidazole, MeI, Na₂CO₃, Pd(OAc)₂, pTosOH, Boc₂O, Et₃N, P₂O₅, HCl H-Gly-OEt, paraformaldehyde, 1,3,5-trioxane, hydrazine monohydrate), Chem-Impex (Wood Dale, USA—Fmoc-Pro-OH, HCl H-β-Ala-OEt), Iris Biotech (Marktredwitz, Germany—Fmoc-Leu-Wang TG resin (0.25 mmol/g)), Novabiochem (Merck, Millipore, Burlington, USA—Oxyma Pure), Acros Organics (Thermo Fisher Scientific, Pittsburgh, USA—picolinic acid), Alfa Aesar (Ward Hill, USA—K₂CO₃). ¹H and ¹³C NMR spectra were recorded at 400 and 101 MHz on a Bruker Avance Neo system (Bruker Corp., Billerica, MA). The chemical shifts were reported in delta (δ) units in parts per million (ppm) with tetramethylsilane as the internal standard. Multiplicities were reported as singlet (s), doublet (d), triplet (t), quartet (q), quintet (qn), multiplet (m), broad (br), or a combination thereof. The coupling constants (J) are given in hertz (Hz). HRMS analysis was recorded with a Micromass Q-TOF-micro system using reserpine (2.10–3 mg/mL, H₂O/MeCN (1:1)) as a reference. Analytical RP-HPLC was performed with Chromolith HR C18 50 × 4.6 column. Semipreparative RP-HPLC-purifications were done using a Gilson system equipped with an INTERCHIM Vydac 150HC C18 column (250 × 22 mm, 10 μm), measured at 215 nm with a flow of 20 mL/min. Automated flash chromatography was performed using a Grace Reveleris X2 equipped with normal phase columns (Grace Silica Flash Cartridges of 40 g). Detection was obtained using an evaporative light scattering detector and a UV detector (214, 254, or 280 nm).

Building Block Synthesis

Peptide Synthesis

All peptides were synthesized manually using the Fmoc/tBu SPPS strategy on the preloaded Fmoc-Leu Wang Tentagel resin (0.25 mmol/g). Fmoc removal was performed by shaking the resin for 5 and then 15 min in the presence of a 20% solution of 4-methylpiperidine in DMF. Tle (3.0 equiv), the first amino acid, was coupled using HBTU/DIPEA (3.0 and 5.0 equiv) for 2 h in DMF. All other standard amino acids were coupled in 3-fold excess using DIC/Oxyma Pure (3.0 and 3.0 equiv) for 2 h in DMF, while nonstandard amino acids (Dmt, d-Arg, constrained dipeptides, and β³hArg) were coupled in 1.5-fold excess for 3 h. Tic(6-OH) was coupled in 1.5-fold excess overnight. All coupling steps were confirmed by Kaiser or Chloranil tests and/or small-scale cleavages were performed and analyzed by LC–MS. Every coupling and deprotecting step was followed by a wash with DMF and CH₂Cl₂ (both three times). All peptides were cleaved with a cleavage cocktail TFA/TIS/H₂O (95:2.5:2.5 v/v/v) at room temperature for 4 h. After evaporation of the cleavage cocktail, the crude was precipitated in cold diethyl ether, followed by lyophilization. The peptides were purified by preparative RP-HPLC using an INTERCHIM Vydac 150HC C18 column (250 × 22 mm, 10 μm) (15–45% gradient in 20 min with MeCN + 0.1% TFA and H₂O + 0.1% TFA as eluent). All hybrids were obtained in purities of >95%.

Competitive Radioligand Binding Assays for MOP and DOP Receptors

Binding assays were performed to evaluate the affinity (K_i) of the different compounds using membrane extracts from HEK293 cells stably expressing hMOP or hDOP. Cells grown to confluence in 150 mm Petri dishes were frozen at −80 °C until use. On the day of the experiment, cells were submitted to a heat shock by placing the Petri dishes at 37 °C for 60 s before returning them to ice. Cells were then harvested using a cell scraper in 50 mM tris-HCl (pH 7.4) and centrifuged at 3200 g for 15 min at 4 °C. The protein concentration was determined with Bio-Rad DC Protein Assay reagents (Bio-Rad Laboratories, Mississauga, ON, Canada), and the pellet was further diluted in 50 mM Tris-HCl buffer (pH 7.4) containing 0.1% bovine serum albumin (BSA) and distributed in 96-well Polypropylene plates (Thermo Fisher Scientific). [¹²⁵I]-deltorphin I (800 Ci/mmol) or [¹²⁵I]-DAMGO (1200 Ci/mmol) was used to determine the binding affinity of each compound in a competitive binding assay on DOP or MOP, respectively. The iodination of deltorphin I and DAMGO was carried out using the iodogen method previously described.⁵³ The K_i values of each compound for DOP and MOP were determined using a membrane concentration of 20 to 40 μg of proteins per well and 1 × 105 counts per minute (cpm) of the radiolabeled ligand. Membranes and the radioligand were incubated for 60 min at room temperature with increasing concentrations of the hybrids ranging from 10 μM to 1 pM for a final sample volume of 200 μL. The reaction was then stopped by rapid vacuum filtration with ice-cold 50 mM tris-HCl (pH 7.4) on 96-well MultiScreen HTS Filter Plate (Millipore Sigma). Filters were then placed in 5 mL tubes, and the radioactivity was determined using the Wizard2 Automatic Gamma Counter (PerkinElmer, Woodbridge, ON, Canada). Data were analyzed using a nonlinear fitting analysis, and the K_i values were calculated using GraphPad Prism 9.0 (GraphPad Software, San Diego, CA) and are expressed as means ± SEM from three independent experiments, each performed in duplicate. The K_i values in the displacement studies were determined from half maximal inhibitory concentration (IC₅₀) values using the Cheng–Prusoff equation with K_d values of 2 nM for the [¹²⁵I]-DAMGO (MOP) and 1.1 nM for the [¹²⁵I]-Deltorphin I (DOP) previously determined from saturation curves.^54,55

Competitive Radioligand Binding Assays for NTS1 and NTS2 Receptors

CHO-K1 cells stably expressing hNTS1 (ES-690-C from PerkinElmer, Montréal, Canada) or 1321N1 cells stably expressing hNTS2 (ES-691-C from PerkinElmer, Montréal, Canada) were cultured in DMEM/F12 culture medium at 37 °C in a humidified chamber under 5% CO₂. Culture media was supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, 20 mM HEPES, and 0.4 mg/mL G418. Cells expressing hNTS1 were frozen when they reached 80% confluency. They were scrapped off the dish with 10 mM Tris-buffer, 1 mM EDTA, pH 7.5 and centrifuged at 15,000g for 5 min at 4 °C. The pellet was then resuspended in 1 mL binding buffer. Cells expressing hNTS2 were also frozen when they reached 80% confluency. As NTS2 is prominently expressed in intracellular compartments, the procedure to isolate hNTS2 differs from hNTS1 isolation.⁵⁶ Cells expressing hNTS2 were scraped off the dish with PBS, 0.5 mM EDTA, pH 7.5, and sonicated for 5 min (pulse 30 s/5 s off, amplitude 40%). They were ultracentrifuged at 100,000g for 60 min at 4 °C. The pellet was resuspended in 1 mL freezing buffer (PBS, glycerol, 0.5 mM EDTA), sonicated for 1 min (pulse 30 s/5 s off, amplitude 50%), and further diluted in binding buffer. Competitive radioligand binding experiments were performed by incubating 15 μg of cell membranes expressing the hNTS1 receptor with 45 pM of ¹²⁵I-[Tyr³]-NT (2200 Ci/mmol) (Revvity, Waltham, Massachusetts, USA) or 50 μg of cell membranes expressing the hNTS2 receptor with 300 pM of ¹²⁵I-[Tyr³]-NT in binding buffer (50 mM Tris-HCl, pH 7.5, 0.2% BSA) in a 96-well Polypropylene plates (Thermo Fisher Scientific). A 60 min incubation at 25 °C in the presence of increasing concentrations of analogs ranging from 100 μM to 10 pM was performed for a final sample volume of 200 μL. Then, the binding reaction mixture was transferred to polyethylenimine (PEI)-coated 96-well filter plates (glass fiber filters GF/B, Millipore, Billerica, MA). The reaction was terminated by filtration using a MultiScreenHTS Vacuum Manifold (Millipore Billerica, MA) and plates were washed three times with 200 μL ice-cold binding buffer. Glass filters were then counted using a γ-counter (2470 Wizard2, PerkinElmer, Mississauga, Ontario, Canada). Nonspecific binding was measured in the presence of 10^–5 M unlabeled NT(8–13) and represented less than 5% of total binding. Data were normalized to NT(8–13) to control for unwanted sources of variation. IC₅₀ values were determined from competition binding curves as the concentration of unlabeled ligand inhibiting 50% of specific ¹²⁵I-[Tyr³]-NT binding. K_i values were determined from IC₅₀ values using the Cheng–Prusoff equation with K_d values of the ¹²⁵I-[Tyr³]-NT previously determined from saturation curves (K_d NTS1 = 0.7 nM and K_d NTS2 = 3.4 nM).^41,54 Competitive radioligand binding data were plotted using the nonlinear regression. One-site-Fit Log(IC₅₀) and represented the mean ± SEM of three independent experiments, each done in triplicate.

BRET Assays for G_αi1 Activation and β-Arrestin-2 Recruitment

BRET-based G_αi1 activation and β-arrestin-2 recruitment assays were used to assess MOP, DOP, and KOP receptor activation. HEK293 cells were cultured in DMEM supplemented with 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 20 mM HEPES in a humidified chamber at 5% CO₂. Twenty-four h before transfection, the cultured HEK293 cells were washed with PBS at room temperature, trypsinized, and seeded at 3 × 106 cells in a 10 cm Petri dish. For the G_αi1 assays, cells were transfected with 4 μg Flag-MOP, Flag-DOP or FLAG-KOP, 800 ng G_αi1-RlucII, 800 ng GFP10-Gγ₁, 400 ng Gβ₂, and 6 μg Salmon Sperm DNA. For the β-arrestin-2 assay, cells were transfected with 1.5 μg hMOP-RlucII (or 2 μg hDOP-RlucII) and 11.5 μg β-arrestin-2-GFP10 (or 11 μg for DOP). The transfection mix was prepared in 600 μL of OptiMeM serum using PEI as a transfection agent at a 3:1 PEI/DNA ratio, as previously described.⁵⁷ The resulting mixture was incubated for 25 min before being added to the cultured cells. Twenty-four h post-transfection, cells were washed with PBS, trypsinized, plated (75,000 cells/well) in 96-well white plates (BD Falcon, Corning, NY), and incubated for a further 24 h. Cells were then equilibrated at 37 °C in a humidified chamber at 5% CO₂ for at least 60 min with 90 μL of HBSS buffer containing 20 mM HEPES. Cells were stimulated with ligand concentrations ranging from a final concentration of 10 μM to 10 pM and incubated for 10 min (β-arrestin-2). The substrate coelenterazine 400A (5 μM) was added and cells were incubated for 5 min prior to the signal acquisition. BRET² signals were acquired using a Mithras⁵⁴ LB 943 Multimode Reader (Berthold, Bad Wildbad, Germany). The BRET² signal was calculated as the ratio of light emitted by the acceptor GFP10 over the light emitted by the donor RLucII. Data were analyzed using GraphPad Prism 9, normalization was done using the BRET² ratio of nonstimulated cells set at 0% and the ratio of cells stimulated with 10 μM DAMGO (for MOP) or 10 μM Deltorphin II (for DOP) set at 100%. EC₅₀ values were obtained using the dose–response stimulation Log (agonist) vs response (three parameters) and represent the mean ± SEM of three independent experiments, each performed in triplicate.

Rat Plasma Stability

Rat plasma was obtained from 2-month-old Sprague–Dawley rats by isolating the translucent phase after centrifuging the blood at 2000g for 15 min in BD Vacutainer K2 EDTA 7.2 mg tube (Becton, Dickinson and Company, NJ, USA). Plasma stability assay was conducted by incubating at 37 °C 27 μL of plasma with 6 μL (1 mM aqueous solution) of the entities for 6, 12, 24, 36, and 48 h (or 1, 2, 5, 10, and 15 min for NT (8–13) and analogs). The enzymatic degradation was stopped by adding 140 μL of 1% formic acid in acetonitrile-ethanol (1:1) solution containing N,N-dimethylbenzamide 0.25 mM (internal standard). The resulting solution was filtered through an Impact Protein Precipitation filter plate (Phenomenex, California) by centrifugation at 8000g for 10 min at 4 °C. After adding 30 μL of distilled water to the filtrates, the solution was analyzed using an Acquity UPLC-MS system class H (column Acquity UPLC CSH C18 (2.1 mm × 50 mm)), 1.7 μm particles with pores 130 AÅ). The method used was 0 to 0.2 min: 5% ACN; 0.2 to 1.5 min: 5 to 95% ACN; 1.5 to 1.8 min: 95% ACN; 1.8 to 2.0 min: 95 to 5% ACN; 2.0 to 2.5 min: 5% ACN in the ESI+ mode. Percentage of remaining compound (not degraded), obtained after MS detection using the MassLynx software (Agilent Technologies, Santa Clara, CA, USA) and corrected to the internal standard (N,N-dimethylbenzamide 0.25 mM), was plotted using GraphPad Prism 9. Half-life was obtained by fitting the one-phase decay and representing the mean ± SEM of three independent experiments.

Glossary

Abbreviations

Aba

4-amino-1,2,4,5-tetrahydro-3H-2-benzazepin-3-one

Ala

alanine

Ana

4-amino-1,2,4,5-tetrahydro-3H-2-naphthoazepin-3-one

Arg

arginine

BRET

bioluminescence resonance energy transfer

BSA

bovine serum albumin

CHO

Chinese hamster ovary

DAMGO

[d-Ala², N-Me-Phe⁴, Gly-ol]-enkephalin

DIC

N,N′-diisopropylcarbodiimide

DIPEA

diisopropylethylamine

DMEM

Dulbecco’s modified eagle medium

Dmt

2,6-dimethyltyrosine

DNA

DNA

DOP

δ-opioid receptor

EDTA

ethylenediaminetetraacetic acid

FBS

fetal bovine serum

FmocOSu

N-(9H-fluoren-9-ylmethoxycarbonyloxy)succinimide

GABA

γ-aminobutyric acid

Gly

glycine

GPCR

G protein-coupled receptor

hArg

homoarginine

HBSS

Hank’s balanced salt solution

HBTU

2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate

HEK

human embryonic kidney

HEPES

4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid

HPLC

high performance liquid chromatography

HRMS

high resolution mass spectrometry

LC-MS

liquid chromatography–mass spectrometry

Leu

leucine

MOP

μ-opioid receptor

MSB

methyl 2-((succinimidooxy)carbonyl)benzoate

Nal

naphthylalanine

NMR

nuclear magnetic resonance

NT

neurotensin

NTS1

neurotensin 1 receptor

NTS2

neurotensin 2 receptor

OPNT

opioid-neurotensin

OUD

opioid use disorder

Oxyma Pure

ethyl cyanohydroxyiminoacetate

PBS

phosphate buffered saline

PEI

polyethylenimine

Phe

phenylalanine

Pro

proline

pTosOH

para-toluene sulfonic acid

RP

reversed phase

SAR

structure–activity relationship

SEM

stander error of the mean

SPPS

solid-phase peptide synthesis

TBDMSCl

tert-butyldimethylsilyl chloride

TBTU

2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium tetrafluoroborate

TFA

trifluoroacetic acid

THF

tetrahydrofuran

Tic

1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid

TIS

triisopropylsilane

TMS

tetramethylsilane

TOF

time-of-flight

Tle

tert-leucine

Tyr

tyrosine

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsptsci.4c00236.

• Data showing the detailed procedures and characterization of the constrained building blocks; analytical characterization of the OPNT hybrids; HPLC chromatograms of the OPNT hybrids and NT-peptides; displacement curves of the OPNT hybrids on hMOP, hDOP, hNTS1, and hNTS2 (PDF)

Author Present Address

^§ Department of Chemical, Biological, Pharmaceutical and Environmental Sciences, University of Messina, Viale Stagno d’Alcontres 31, 98166 Messina, Italy

Author Contributions

^∥ Co-first authors J.D.N. and É.B. contributed equally. The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. J.D.N. performed synthesis of the building blocks as well as the peptides, analyzed the in vitro results, and contributed to the writing. É.B. performed the in vitro experiments, analyzed the results, and contributed to the writing. S.P. and E.V. performed synthesis. B.L., M.C., R.B., A.L., and B.J.H. helped with the in vitro experiments. D.T., J-M.L., L.G., P.S., and S.B. designed experiments, analyzed data, contributed to the writing, and oversaw the study.

J.D.N. and S.B. were supported by the Research Foundation Flanders (FWO Vlaanderen; grant number 1SB0422N) and S.B. acknowledges the Research Council of the VUB for support through the Strategic Research Programme (SRP50 & SRP95) and FWO Hercules (OZR3584; OZR3939). É.B., R.B. and A.L. were supported by research scholarships awarded by the Fonds de recherche du Québec – Santé (FRQS). É.B. and R.B. also acknowledge the Canadian Institutes of Health Research (CIHR) for support through research scholarships. This work was supported by a Canadian Institutes of Health Research (CIHR) grant (FDN-148413) to PS and a Natural Sciences and Engineering Research Council of Canada (NSERC) grant to LG. PS holds a Canada Research Chair in Neurophysiopharmacology of Chronic Pain.

The authors declare no competing financial interest.