The peptide nociceptin/orphanin FQ (N/OFQ) is the endogenous ligand of the N/OFQ peptide (NOP) receptor (Meunier et al., 1995; Reinscheid et al., 1995). The N/OFQ-NOP receptor system regulates several biological functions, including pain transmission, locomotor activity, memory, emotional states, food intake, drug abuse, micturition, cough reflexes, cardiovascular, respiratory, and immune functions (Lambert, 2008; Toll et al., 2016). The NOP receptor represents an innovative pharmacological target for the treatment of several conditions, including pain (Schröder et al., 2014; Calo and Lambert, 2018), depression (Gavioli and Calo’, 2013), drug addiction (Ciccocioppo et al., 2019), Parkinson’s disease (Mercatelli et al., 2020), and incontinence due to overactive bladder (Angelico et al., 2019). However, the pleiotropic effects exerted by N/OFQ and NOP ligands may limit their drug development. For instance, the sedative effects associated with high doses of NOP agonists hampered their development as antitussives

(Woodcock et al., 2010) or analgesics (Byford et al., 2007). We hypothesized that one way to trigger some of the NOP mediated effects instead of others, might be the so-called functional selectivity of NOP agonists.

Functional selectivity, or biased agonism, is the ability of a ligand to stabilize different active conformations of the same receptor and selectively activate different downstream signaling pathways (i.e., promoting the interaction of the receptor with G protein over arrestin recruitment or vice versa). Biased agonists have been proposed as “smarter drugs,” with the potential of specifically targeting therapeutic signaling pathways while avoiding others that could lead to side effects (Kenakin, 2019). In recent years biased ligands have been identified and characterized for several different G protein coupled receptors (Tan et al., 2018); however, despite the high potential of the biased agonism strategy in drug discovery, we still need to substantially increase our knowledge regarding translation of in vitro bias to in vivo drug effects (Michel and Charlton, 2018; Kenakin, 2019). With the NOP receptor, in vitro studies have been performed to investigate the ability of NOP ligands to differently recruit G protein and β-arrestin (Chang et al., 2015; Malfacini et al., 2015; Ferrari et al., 2016, 2017; Pacifico et al., 2020). Additionally, there is in vitro and in vivo evidence that NOP agonists are able to differently activate G protein-dependent events and receptor phosphorylation/and internalization (Mann et al., 2019). As far as in vivo impact of NOP functional selectivity is concerned, the only information available suggests that NOP ligands producing similar effects on NOP/G protein interaction but showing different effects on β-arrestin 2 recruitment, elicited different actions on anxiety and mood (Asth et al., 2016).

The aim of the present study was to investigate the in vivo impact of NOP receptor functional selectivity; in particular we tested the hypothesis that NOP biased agonism may be instrumental for identifying effective analgesics devoid of the locomotor impairing effects associated with NOP activation. To this aim we integrated the results of in vitro, and genetic and pharmacological in vivo studies: (i) NOP selective agonists (Ro 65-6570, AT-403, and MCOPPB) characterized by different degree of biased agonism toward G protein (Ferrari et al., 2017) were studied and compared in the dynamic mass redistribution (DMR) test, a label-free assay recently validated for the NOP receptor in our laboratories (Malfacini et al., 2018), (ii) the antinociceptive and motor effects of Ro 65-6570 were evaluated in wild type and in mice knockout for the β-arrestin2 gene [βarr2(−/−)], and iii) dose response curves for antinociceptive and motor effects were performed in normal mice using Ro 65-6570, AT-403, and MCOPPB in order to estimate their therapeutic index. The formalin assay was used as analgesiometric test while locomotor activity and rotarod assays were used to determine locomotor and coordination impairment. The chemical structure of Ro 65-6570, AT-403, and MCOPPB is shown in Figure 1. These NOP agonists have been selected based on their different degree of biased agonism; in fact, a previous study performed in our laboratories demonstrated that AT-403 behaves as an unbiased NOP agonist while Ro 65-6570 and MCOPPB are biased toward G protein with bias factors of 1.64 and 0.97, respectively (Ferrari et al., 2017).

This study investigated the possible impact of NOP receptor functional selectivity on antinociceptive vs. motor impairing effects of NOP agonists in vivo. βarr2(−/−) mice displayed slightly increased antinociceptive effects associated with reduced motor impairment in response to the NOP agonist Ro 65-6570. The hypothesis that NOP agonists biased toward G protein might display a larger therapeutic index was tested by comparing the antinociceptive and motor impairing potencies of Ro 65-6570, AT-403, and MCOPPB in wild type mice. Despite 30-fold differences in their bias factor values, NOP agonists displayed a similar therapeutic index (approximately 10) between their antinociceptive vs. hypolocomotor effects.

To investigate the role of βarr2 in the antinociceptive and locomotor inhibiting effects of NOP agonists, experiments were performed with βarr2(−/−) mice and the NOP selective agonist Ro 65-6570. This compound, identified in Roche laboratories (Wichmann et al., 1999), has been used in several in vitro and in vivo pharmacological studies (Toll et al., 2016). Importantly the NOP selectivity of Ro 65-6570 in vivo has been previously demonstrated both in NOP antagonism and knockout mice studies (Byford et al., 2007; Rutten et al., 2010; Asth et al., 2016; Schiene et al., 2018).

No phenotype differences were detected between βarr2(+/+) and βarr2(−/−) mice with respect to their sensitivity to formalin and motor performance on the rotarod. This confirms previous studies showing no differences in terms of nociception between βarr2(+/+) and βarr2(−/−) mice (Bohn et al., 1999; Mittal et al., 2012). Additionally, no differences were reported in the literature between the basal locomotor activity of βarr2(+/+) and βarr2(−/−) mice (Beaulieu et al., 2005; Urs et al., 2011; Del’Guidice and Beaulieu, 2015; Toth et al., 2018). βarr2(-/-) mice showed different sensitivity to the antinociceptive and motor impairing effects of Ro 65-6570. Specifically, Ro 65-6570 1 mg/kg failed to produce antinociceptive effects in βarr2(+/+) but was effective in βarr2(−/−) mice. The lack of effect of Ro 65-6570 1 mg/kg in wild-type mice can be ascribed to the fact that C57BL/6J mice are less sensitive to the antinociceptive effects of NOP agonists than CD-1 mice used in our previous studies (Rizzi et al., 2016). Higher doses of Ro 65-6570 could not be tested in the formalin assay because of their motor inhibiting effects. Mice lacking the βarr2 protein were sensitive to the antinociceptive effects of 1 mg/kg Ro 65-6570. This result parallels findings obtained with morphine whose antinociceptive action is potentiated in βarr2(−/−) mice (Bohn et al., 1999; Azevedo Neto et al., 2020) as well as in phosphorylation-deficient mu knock-in mice (Kliewer et al., 2019). Collectively these results can be interpreted assuming that βarr2 acts [as demonstrated for several GPCRs (Gurevich and Gurevich, 2019)] as a desensitizing element of the antinociceptive response to mu opioid as well as NOP receptor agonists. Of note, NOP receptor phosphorylation and internalization have already been described by different research groups (Spampinato et al., 2001; Corbani et al., 2004; Zhang et al., 2012; Bird et al., 2018; Mann et al., 2019). βarr2(−/−) mice were slightly less sensitive than βarr2(+/+) mice to the effects of Ro 65-6570 in the rotarod, with the 3 mg/kg dose being active in wild-type animals but not in mice lacking the βarr2 gene. Thus, the βarr2 protein seems to contribute, at least in part, to mediating the inhibitory effect of NOP agonists on locomotion; however, to the best of our knowledge no direct evidence in support of this suggestion is available in the literature. It is worthy of mention that βarr2(−/−) mice are reported to be less sensitive to the effects of drugs altering locomotion. This has been demonstrated both for stimulants [i.e., amphetamine, apomorphine (Beaulieu et al., 2005), and morphine (Urs et al., 2011)] and for drugs reducing locomotor activity [i.e., lithium (Beaulieu et al., 2008; Del’Guidice and Beaulieu, 2015), lamotrigine and valproate (Del’Guidice and Beaulieu, 2015), and the growth hormone secretagogue receptor 1a antagonist YIL781 (Toth et al., 2018)]. The authors of these studies ascribed this finding to the involvement of βarr2 in the intracellular events that follow dopamine receptor activation.

The experiment with βarr2(−/−) mice prompted us to hypothesize that NOP agonist biased toward G protein may display a larger therapeutic window in terms of analgesia vs. motor impairment side effects. To test this attractive hypothesis we used three chemically different NOP agonists Ro 65-6570, AT-403, and MCOPPB. DMR experiments performed under the experimental conditions previously described by Malfacini et al. (2018) demonstrated that these compounds act as NOP full agonists with the following rank order of potency MCOPPB > AT-403 > Ro 65-6570. These results are perfectly in line with findings previously obtained in different pharmacological assays including stimulation of [³⁵S]GTPγS [binding, calcium mobilization, BRET assay measuring NOP interaction with G protein and βarr2, and the electrically stimulated mouse vas deferens bioassay (Ferrari et al., 2017). Importantly, BRET studies demonstrated that these NOP agonists displayed different G protein vs. βarr2 bias factors; in particular AT-403 (0.16) behaved as an unbiased agonist while MCOPPB (0.97) and Ro 65-6570 (1.64) were biased toward G protein (Ferrari et al., 2017). Similar values of bias factor have been previously reported for the latter compounds (Chang et al., 2015; Malfacini et al., 2015).

In mice MCOPPB and AT-403 produced antinociceptive effects in the formalin assay, being active at 1 and 0.1 mg/kg, respectively. A similar antinociceptive effect was recorded for Ro 65-6570 in the formalin assay at the dose of 1 mg/kg (Rizzi et al., 2016). Of note, the second phase of the assay seems more sensitive to the effect of NOP agonists, suggesting an involvement of the NOP receptor mainly in regulating the inflammatory phase of the test. This hypothesis is corroborated by the fact that NOP antagonists produced pro-nociceptive effects only in the second phase of the formalin test (Rizzi et al., 2006). However, the naturally occurring NOP agonist N/OFQ, after intrathecal administration, was able to reduce both the first and the second phase of the rat formalin assay (Yamamoto et al., 1997, 2000). All the NOP agonists tested produced antinociceptive effects, with the following rank order of potency: AT-403 > MCOPPB = Ro 65-6570. Of note the in vivo potency of MCOPPB is relatively low; this is, however, in line with previous studies that investigated its actions on anxiety (Hirao et al., 2008) and mood (Holanda et al., 2019). The low potency of MCOPPB in vivo can be possibly ascribed to pharmacokinetic issues. As far as the reduction of locomotor activity is concerned, both AT-403 and Ro 65-6570 significantly reduced mouse locomotion at 1 and 10 mg/kg doses, respectively. This effect is due to the selective activation of the NOP receptor, since it completely disappeared in NOP(−/−) mice. Differently, MCOPPB failed to produce significant effects on locomotor activity up to 10 mg/kg, in line with previous findings (Hirao et al., 2008). The reason for this lack of effect of MCOPPB on locomotion is not known. We can speculate that, despite the NOP occupancy higher than 50% measured in the whole mouse brain (Hayashi et al., 2009), at 10 mg/kg i.p. this compound is not able to reach, at least in sufficient concentrations, those brain areas important for the hypolocomotor effects of NOP agonists (Marti et al., 2004, 2009; Narayanan et al., 2004; Byford et al., 2007).

What is clear from the present experiments is that NOP agonists displayed the same therapeutic index with hypolocomotor doses being approximately 10-fold higher than antinociceptive doses. In particular, no differences were detected between the unbiased agonist AT-403 and the G protein biased agonist Ro 65-6570, suggesting that the different ability of inducing G protein vs. βarr2 recruitment of a NOP agonist cannot be applied to predict the antinociceptive vs. hypolocomotor properties. These pharmacological findings fail to support the hypothesis that NOP receptor functional selectivity (i.e., G protein vs. βarr2 coupling) can be a useful strategy to obtain better tolerated analgesics, however some caution should be adopted in the interpretation of these results. In fact, the ligands used in this study display bias factor values of 0.94 (MCOPPB) and 1.64 (Ro 65-6570), meaning that the difference in the recruitment of G protein vs. βarr2 is of 10 and 40-fold, respectively. Therefore, it is possible that these values of bias factor are too low to have significant impact in vivo. For instance, it has been reported that MCOPPB and the Ro 65-6570 analog Ro 64-6198, are able to induce NOP receptor phosphorylation and internalization similar to the endogenous NOP agonist N/OFQ, both in vitro and in vivo (Mann et al., 2019), demonstrating that βarr2 dependent biological events can occur after the treatment with these NOP ligands. It is possibly worthy of mention that the NOP agonist cebranopadol displays a very large bias toward G protein in vitro and acts in vivo as a potent antinociceptive drug devoid of inhibitory actions on mouse motor performance (Rizzi et al., 2016). However, the in vivo pharmacological actions of cebranopadol cannot be exclusively attributed to its NOP agonist activity since this molecule is able to simultaneously activate NOP and classical opioid receptors (Calo and Lambert, 2018; Tzschentke et al., 2019). Unfortunately, nothing is known about the functional selectivity of these compounds for the different types of inhibitory G proteins that can be activated by the NOP receptor, this notion could help to explain a possible significance of biased agonism for the NOP receptor.

Finally we would like to underline some limitations of the present study. First, the estimate of the NOP agonist therapeutic indexes is far from being optimal since some ED₅₀ values could be only roughly estimated. Second, some comparisons between set of data have to be looked at with caution because different strains (C57BL/6J vs. CD-1) of mice and drug route of administration (i.p. vs. i.v.) have been used. Third, no information is available on the pharmacokinetic properties of the drugs used in this study; such information could significantly contribute to correctly interpret in vivo findings. However, at least in our opinion, the above mentioned limitations do not substantially compromise the major finding of the study that is lack of correlation between in vitro biased agonism and in vivo therapeutic index of NOP agonists.

In conclusion, this study investigated the possible impact of functional selectivity (G protein vs. βarr2) on the ability of NOP agonists to promote analgesic vs. motor impairing effects. Studies with βarr2(−/−) mice suggested that bias toward G protein might be associated with a wider therapeutic index, however, this hypothesis was not corroborated by pharmacological studies that demonstrated how unbiased and G protein biased NOP agonists display similar antinociceptive vs. motor impairing properties.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

The animal study was reviewed and approved by the Animal Welfare Body of the University of Ferrara and Italian Ministry of Health (License N° 302/2017).

CR and GC designed this study. CR, JA, CS, and DM performed the experiments and conducted the statistical analysis. SP and NZ synthesized Ro 65-6570 and AT-403, respectively. CR and GC prepared the manuscript. JA, CR, CS, DM, SP, NZ, and GC revised the manuscript into its finalized form. All authors contributed to the article and approved the submitted version.

NZ was an employee of Astraea Therapeutics. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.