In vivo antinociception of potent mu opioid agonist tetrapeptide analogues and comparison with a compact opioid agonist - neurokinin 1 receptor antagonist chimera

Opioids are the most widely used analgesics in the treatment of various pain states. Early binding studies and functional bioassays defined three main types of opioid receptors in the central nervous system: μ- (MOR), δ- (DOR) and κ- (KOR) receptors [1, 2]. The μ-opioid receptor was identified to be essential for an efficient antinociception in acute and severe pain models. Aside from the desired pain-relieving action, prolonged exposure to μ opioids results in well established, undesired side effects, including sedation, respiratory depression, physical dependence and analgesic tolerance [3, 4]. Several studies however provided evidence that compounds with a dual MOR/DOR activity present beneficial pharmacological effects in comparison to highly selective MOR agonists [5]. It was, for example, demonstrated that the analgesic efficacy of MOR agonists can be enhanced by DOR agonists [6, 7]. Moreover, mixed MOR agonism/DOR (ant)agonism can suppress or eliminate the development of physical dependence, tolerance and respiratory depression, adverse effects which are typically observed for selective MOR agonists [5, 8, 9]. Because of these promising indications, compounds which combine MOR agonism with DOR (ant)agonism have received a great deal of attention as lead structures towards novel opioid analgesics.

A limitation of most peptides for use as analgesics consists of their limited access to the brain. This drawback is not only due to their poor metabolic stability, but also a consequence of their low lipid solubility and limited ability to cross the blood-brain barrier (BBB) [10]. Several strategies have been developed to overcome the hurdle of opioid peptide drug delivery to the brain, including structural modifications such as glycosylation [10, 11], lipid conjugation [12], and various types of cyclization [13‐15]. In this way researchers attempt to modulate the balance of hydrophobic-hydrophilic properties of the compounds and to decrease the number of accessible conformations, which can potentially result in an increased membrane permeability. Despite the preparation and evaluation of a plethora of bioactive opioid ligands, both peptidic and non-peptidic, an alternative to morphine, the golden standard in moderate to severe pain research, has yet to be found.

An alternative to the preparation of compounds which solely target the opioid system consists of the creation of hybrid compounds that interact with multiple biologically relevant targets [16‐18]. The emerging strategy to design and synthesize chimeras, also called designed multiple ligands (DMLs), as a promising alternative to a treatment by monotherapy, is justified by the fact that such compounds are commonly characterized by not only a high activity, but more importantly present the possibility to improve the pharmacological profile by selectively 'designing out' an undesired biological characteristic, both in terms of action or transport [19, 20]. As such, the chemical structure of the drug allows to modulate the permeability of the active substance and in consequence create "site specificity of action" [9, 16]. The precision of action of these kinds of compounds can be built up by the knowledge about their specific activity as single molecules. By hybridizing two active compounds into one chemical entity it is possible to develop drugs which may provide a superior delivery capacity, as compared to the delivery of a physical mixture of the drugs.

When applied to pain research the combination of an opioid pharmacophore and a NK1 receptor antagonist led to a new type of bifunctional drugs which aim at a prolonged therapeutic efficiency over time [21‐24]. Substance P, an 11-amino acid peptide belonging to the tachykinin family, is known to be important for transmission of nociceptive signals [25]. The design rationale of the opioid-NK1 DMLs originates from the upregulation of the pronociceptive neurotransmitter substance P and its receptor, the neurokinin 1 receptor (NK1R), upon chronic administration of opioids [26, 27]. The NK1 antagonist pharmacophore in these DMLs could potentially counter the enhanced expression of substance P and its preferential receptor, NK1R.

From our previous results, we selected three opioid tetrapeptides with promising in vitro potency for further in vivo evaluation. All three reported peptide analogues were derived from dermorphin (H-Tyr-D-Ala-Phe-Gly-Tyr-Pro-Ser-NH₂) as this exogenous peptide proved to have a high affinity and selectivity for the MOR [28, 29]. Additionally, its use as a lead structure was also motivated by the fact that dermorphin, when administered centrally or peripherally in rodents, induces less tolerance development when compared to morphine [30]. We recently identified the opioid ligands H-Dmt-NMe-D-Ala-Phe-Sar-NH₂ (BVD02) [31], H-Dmt-NMe-D-Ala-Aba-Gly-NH₂ (BVD03 with Aba: 4-amino-1,2,4,5-tetrahydro-2-benzazepin-3-one, see Figure 1) [31] and H-Dmt-D-Arg-Aba-Gly-NH₂ (AN81) [32] to be potent compounds in vitro (Figure 1 and Table 1). The N-terminal Tyr¹ present in dermorphin was replaced by 2',6'-dimethyl-(S)-tyrosine (Dmt) to improve the enzymatic stability, receptor affinity and opioid potency [33], whereas a conformationally constrained aminobenzazepinone (Aba) moiety placed in third position maintained MOR affinity as well as activation and resulted in a favorable increase in DOR binding and activity [34‐36].

The Aba benzazepine moiety was also used as a central scaffold in the design of a compact bifunctional ligand, SBCHM01, which possessed the desired dual opioid agonist - NK1R antagonist activity in vitro (Figure 1 and Table 1) [32].

The potent antinociceptive effects of the investigated structures after peripheral administration indicate that all four peptidic compounds are able to cross the BBB. The antinociceptive enhancement concomitant with the introduction of a conformational constraint at the level of the Phe³ residue (BVD02 vs. BVD03) indicates that next to molecular lipophilicity, an additional factor for efficient analgesia in these peptidic ligands consists of the limitation in conformational flexibility. The observed effects are in agreement with earlier in vitro binding data obtained for these peptides (BVD02→BVD03: K_iμ 60 → 15 nM, K_iδ 130 → 5 nM, Table 1) [31], which also confirm that the conformational restriction imposed by the Aba moiety maintains the MOR binding affinity of the ligands, but substantially ameliorates DOR binding. The receptor affinities were in agreement with functional in vitro MOR and DOR activity, as determined by a functional forskolin-stimulated cAMP accumulation assay (BVD02→BVD03: EC₅₀μ 0.079 → 0.00174 nM, EC₅₀δ 4400 → 0.016 nM, Table 1) [31]. The aforementioned role of delta receptors in the modulation of efficacy of MOR agonists needs to be considered at this point. Next to the MOR activation needed for efficacious analgesia, the superior MOR binding and potency of BVD03 in comparison to BVD02 (factor of 4 and 45, respectively), is accompanied by a 26-fold increase in DOR affinity and a 10⁴ gain in DOR potency. Given the proposed synergy of mixed MOR agonists-DOR agonists in analgesia, it is likely that both MOR and DOR components of BVD03 are responsible for the enhanced in vivo antinociception when compared to BVD02. The activity data suggest that next to the N-methylation in the "ring opened" BVD02, leading to an increased lipophilicity and potentially higher bioavailability, an additional constraint of the Phe side chain by use of Aba in BVD03 is also beneficial for antinociception (Figure 2A and 2B).

In an earlier study, and based on the peptide sequence of the lead MOR agonist Dmt¹-DALDA (H-Dmt-D-Arg-Phe-Lys-NH₂) [42], which is a potent opioid analgesic [18], we prepared and evaluated the in vitro pharmacological opioid profile of H-Dmt-D-Arg-Aba-Gly-NH₂ (AN81)_. This compound was also identified as a full MOR/DOR agonist with subnanomolar affinity and tissue bioactivity (Table 1) [32]. The structure differs from BVD03 only in the amide nitrogen N-methyl group and side chain of the second residue in the peptide sequence (Figure 1). When an activity comparison is made between the two peptides at different doses (see Figure 2B and 2C), AN81 seems to produce a faster build-up in antinociceptive response over time at a dose of 4 mg/kg, which also seems to be maintained during the first two hours of measurements. From Figure 3, we concluded that at that dose of 4 mg/kg AN81 is superior to BVD03 and furthermore possesses a very high %MPE value even after two hours. AN81's remarkable antinociceptive potency is observed at doses ranging from 4 mg/kg (i.v.) to 0.5 mg/kg (i.v.) (Figure 2C), showing that by lowering the dose there is the possibility to obtain the same long-lasting pain-relieving effect. To have a better idea of the total duration of action of the best tetrapeptide analogue (AN81), an additional in vivo pain test was used. We switched to the radiant tail-flick assay to avoid the infliction of a possible tail injury after repetitive measurements in the hot-water test. Such an injury could in turn result in the erroneous measurements of antinociception as observed in earlier studies by our group upon chronic administration of potent drugs.

A subcutaneous (s.c.) administration of AN81 at three different doses (1, 0.5 and 0.2 mg/kg) induced marked analgesia lasting up to 7 hours, whereas morphine lost almost all activity over a time period of 2 hours at more elevated doses, relative to AN81 (Figure 4). Because of the in vivo potency and long duration of action, these results place AN81 on the list of potentially promising active lead compounds for acute pain treatment.

The ability of the two most active compounds AN81 and BVD03 to circumvent one of morphine's major and highly problematic side effects, the development of tolerance upon chronic opioid administration was subsequently tested. Figure 3 clearly demonstrates that a daily dose of both peptide ligands generate an important reduction in antinociception of the fifth day of the experiment. Only the most active derivative, AN81, showed a partially preserved activity with a maximum (ca. 35%MPE) at the time point of 30 min post-injection. An observation which can potentially be correlated to the balanced MOR/DOR agonism of AN81.

The combination of an opioid pharmacophore and a NK1 receptor antagonist in one chemical entity has been identified as an approach towards a new type of bifunctional drugs which aim at a prolonged therapeutic efficiency over time [21‐24]. When a 3',5'-bistrifluoromethyl benzyl moiety is coupled to the C-terminus of AN81, a compound with dual opioid - NK1 in vitro activity results (Table 1) [32]. After systemic injection, SBCHM01 produced a significant dose- and time-dependent analgesia during 2 to 3 hours (Figure 5 and 6, resp.). The inferior antinociceptive effect of this chimera, when compared to its 'pure' opioid analogue AN81 (Figure 2C and 2D), could not only be ascribed to a diminished MOR affinity and activation (Table 1), but is also in agreement with the large decrease in DOR binding and potency (17- and 103-fold, resp.). The results obtained for SBCHM01 indicate that neither an synergistic nor potentiating antinociceptive effect appears when both pharmacophores are combined in a single chemical entity.

The opioid-induced development of tolerance observed with BVD03 and AN81 (Figure 3), was in this study unfortunately not reduced by the additional presence of a NK1R antagonist pharmacophoric group covalently attached to the C-terminus of the opioid tetrapeptide AN81 (Figures 5 and 6). Concomitant with the substantial loss of DOR agonism in SBCHM01, the beneficial property of dual MOR/DOR agonists in lowering opioid-induced tolerance is unfortunately eliminated, and could explain the complete loss in antinociception on the fifth day of the experiment. This in contrast to the partial preservation of analgesia observed after repetitive administration of AN81, a balanced and potent MOR/DOR agonist. (Figure 3).

Recently, Vanderah and coworkers [43] reported the detailed in vivo evaluation of H-Tyr-D-Ala-Gly-Phe-Met-Pro-Leu-Trp-O-3',5'-Bn(CF₃)₂ (TY005) in several animal assays to define the role of this compound in thermal and tactile stimuli in uninjured, sham- and nerve-injured male, Sprague-Dawley rats. They observed that this compound was able to exert the desired dual activity in vivo. The opioid agonist activity and NK1R antagonism were demonstrated in independent assays in which one of the two activities was isolated by blocking the second functionality. The authors showed that the development of antihyperalgesic tolerance could be suppressed by this multimodal ligand. This compound was further optimized by Hruby and coworkers by introducing a Dmt residue in position 1 and replacing the metabolically labile ester function by an amide group and demonstrated that these changes resulted in improved opioid agonism, while maintaining NK1 activity [24].

Upon comparison of this study with related work by other research teams [43, 44], it is however important to point out that different animal models and conditions (animals, assay type, drug delivery) were applied. In the recent publication of Largent-Milnes et al. [43] the inhibition of antihyperalgesic tolerance with the hybrid opioid-NK1 TY005, was demonstrated after i.t. administration and by use of the paw withdrawal test. This is in contrast to the current study, where the hot water tail-flick assay served as a model for acute pain and in which the drug was injected intravenously. The various types of noxious stimuli employed in the experimental paradigms, as well as the different administration routes, make it difficult to rationalize the different mechanisms that are responsible of this apparent discrepancy.

In the present study AN81 (H-Dmt-D-Arg-Aba-Gly-NH₂) exhibits the strongest analgesic effect, when administered systemically, which suggests that this peptidomimetic has the capacity to penetrate the highly selective blood brain barrier. Moreover, this opioid tetrapeptide analogue is characterized by a long duration of action after subcutaneous injection and proved to be more active than a very similar structure with a NMe-D-Ala residue in the second position of the sequence, i.e. BVD03. This conformationally restricted peptide ligand proved in turn to be more potent that its "ring opened" analogue BVD02.

Unfortunately, AN81 appeared to still induce the development of tolerance after repetitive administration over a time period of five days when tested in the hot water tail-flick test. When compared to morphine, the induced tolerance was however markedly lower. A similar result was obtained for the bifunctional opioid agonist-NK1R antagonist ligand, SBCHM01. This peptidic chimera was, in analogy with the 'pure' opioid ligands, capable of reaching and activating opioid receptors in the CNS, but it lost its antinociceptive potency completely upon chronic injection as determined in the hot water tail-flick assay.

The apparent inability of SBCHM01 to suppress tolerance development appears to oppose recent reports by Vanderah and coworkers [43] which indicate that a hybrid opioid-NK1 octapeptide ligand (TY005) was able to attenuate tolerance development related to sustained opioid pathophysiology in an hyperalgesic model after central (i.t.) administration, as opposed to this study in which i.v. administration was used in an acute pain model. For a general clinical applicability, an important goal in pain research consists of the discovery and development of improved analgesic drugs which could be administered systemically. Hence, we are convinced that unravelling the cause of this intriguing discrepancy is important and will consequently be the target of our future SAR studies.