Oxytocin and Vasopressin Agonists and Antagonists as Research Tools and Potential Therapeutics

We recently reviewed the status of peptide and nonpeptide agonists and antagonists for the V1a, V1b and V2 receptors for arginine vasopressin (AVP) and the oxytocin receptor for oxytocin (OT). In the present review, we update the status of peptides and nonpeptides as: (i) research tools and (ii) therapeutic agents. We also present our recent findings on the design of fluorescent ligands for V1b receptor localisation and for OT receptor dimerisation. We note the exciting discoveries regarding two novel naturally occurring analogues of OT. Recent reports of a selective VP V1a agonist and a selective OT agonist point to the continued therapeutic potential of peptides in this field. To date, only two nonpeptides, the V2/V1a antagonist, conivaptan and the V2 antagonist tolvaptan have received Food and Drug Administration approval for clinical use. The development of nonpeptide AVP V1a, V1b and V2 antagonists and OT agonists and antagonists has recently been abandoned by Merck, Sanofi and Pfizer. A promising OT antagonist, Retosiban, developed at Glaxo SmithKline is currently in a Phase II clinical trial for the prevention of premature labour. A number of the nonpeptide ligands that were not successful in clinical trials are proving to be valuable as research tools. Peptide agonists and antagonists continue to be very widely used as research tools in this field. In this regard, we present receptor data on some of the most widely used peptide and nonpeptide ligands, as a guide for their use, especially with regard to receptor selectivity and species differences.

Arginine vasopressin mediates its actions through three known receptors: V 1a , V 1b and V 2 . V 1a receptors are expressed in the liver, vascular smooth muscle cells, brain and in many other tissues (12,21,22). In the vasculature, V 1a receptors mediate the pressor actions of AVP by a phospholipase C-mediated pathway. In the brain, V 1a receptors mediate the anxiety producing responses to AVP (27,36). V 1b receptors, discovered long after the V 1a and V 2 receptors, present in the anterior pituitary, mediate the adrenocorticotrophic hormone-releasing effects of AVP, also by a phospholipase C-mediated pathway (22) Evidence for the presence of V 1b receptors in extra-pituitary tissues such as brain, the kidney and the adrenal medulla has also been reported (37). Recently, the V 1b receptor has been shown to mediate anxiety and stress in rats and in humans (45). V 2 receptors, present in the collecting duct of the kidney, mediate the antidiuretic action of AVP by an adenylate cyclase-mediated pathway (12,21,22).

Scope of the present review
We have previously reviewed the status of developments in the design and synthesis of peptide and nonpeptide AVP and OT agonists and antagonists (1). Here, we focus on the properties of the most widely used peptides requested from the Manning laboratory or purchased from suppliers, together with some recently reported potential clinically useful peptides from the Ferring Laboratory (38,39). Space considerations preclude our being able to present or to discuss recent synthetic studies carried out in other laboratories (40)(41)(42)(43). We also update the current status of the pre-clinical and clinical development of nonpeptide AVP and OT antagonists and of the pre-clinical development of nonpeptide OT agonists (44). The excellent reviews on nonpeptide AVP antagonists (45) and on nonpeptide OT antagonists (46) should be consulted for more in-depth presentations of their chemistry and pharmacology. We also review the merits of peptide versus nonpeptide AVP and OT agonists and antagonists as: (i) research tools and (ii) therapeutic agents. We present human and rat receptor data for a number of selective peptide agonists and for both peptide and nonpeptide antagonists. We illustrate the need to be aware of: (i) species differences, (ii) selectivity differences and (iii) in vitro-in vivo differences when using a specific ligand for receptor characterisation. Finally, we present the highlights of our recent studies aimed at: (i) the development of selective fluorescent ligands for the rat and human V 1b receptors (47) and (ii) the development of fluorescence based strategies that have been used to prove the existence of OT receptor dimers in native tissue (48).

Peptide synthesis
All the OT and AVP agonists, antagonists, radiolabelled and fluorescent ligands from our laboratories were synthesised using the Merrifield solid-phase method (4,49). The synthetic strategy relies very heavily on methodology developed in the du Vigneaud laboratory for the original syntheses of OT and AVP (2,3). The procedures used are described in the original publications cited here. For other references, see Manning et al. (1).

Bioassays
All of the published peptides from our laboratories, presented in Tables 1, 3-8, were assayed for agonistic and antagonistic activities in in vitro and in vivo rat oxytocic assays, in the rat vasopressor assay and in the rat antidiuretic assay in the laboratories of our collaborators Drs Wilbur H. Sawyer, W. Y. Chan and Hazel Szeto. For agonists, the four-point assay design (50) was used and for antagonists, Schild's pA 2 method (51) was employed. The pA 2 is the negative logarithm of the molar concentration of the antagonist that requires a two-fold increase in agonist concentration to achieve the same effect as that found in the absence of antagonist. In practice, this concentration is estimated by finding concentrations above and below the pA 2 dose and interpolating on a logarithmic scale.
In the rat in vivo assays, the pA 2 (effective dose) is divided by an arbitrarily assumed volume of distribution of 67 ml ⁄ kg (52) in an attempt to derive the approximate molar concentration [M] of the pA 2 dose in the vicinity of the receptors. Thus, in vivo pA 2 values are very approximate estimates. The USP Posterior Pituitary Reference Standard or synthetic OT and AVP, which had been standardised in oxytocic and vasopressor units against this standard, were used as agonists for working standards in all bioassays. In vitro oxytocic assays were performed on isolated uteri from diethylstilbestrol-primed rats in a Mg 2+ -free van Dyke Hasting's solution (53). In vivo anti-OT potencies were determined in urethane-anaesthetised diethylstilbestrol-primed rats as described previously (54,55). Vasopressor assays were performed on urethane-anaesthetised and phenoxybenzamine-treated rats as described by Dekanski (55), and antidiuretic assays on water-loaded rats under ethanol anesthesia as described by Sawyer (56).
The two new OT-related analogues given in Table 2 (1, 2) have not yet been evaluated in standard rat bioassays. The discovery of Selective vasopressin V 2 receptor agonists (Table 3) AVP is equipotent as an antidiuretic agonist and as a vasopressor agonist (79) (Table 3). Thus, it is totally nonselective. It is also not selective with respect to its oxytocic activity. The three analogues of AVP, peptides 1-3 in Table 3 namely; dDAVP, VDAVP and dVDAVP, exhibit striking gains in antidiuretic ⁄ vasopressor selectivity. All three peptides have been widely used as selective V 2 agonists. dDAVP, first synthesised by the Zaoral et al. (80) in Prague and later licensed to Ferring, has long been the drug of choice for the treatment of diabetes insipidus. It has been marketed under the trademark Desmopressin (Minirin). The human receptor affinities for dDAVP and dVDAVP given in Table 11 shows clearly that dVDAVP has a ten-fold higher affinity for the human VP V 2 receptor than dDAVP. However, both peptides also exhibit high affinities for the human V 1b receptor and to somewhat lesser extent for the human OT receptor (Table  11). So clearly they are not selective V 2 agonists in humans with respect to both hV 1b or hV 1a receptors. The search for a V 2 agonist that is selective with respect to the V 1a and V 1b receptors in humans is still a challenging goal in this field. Yet, in the rat, dDAVP could be considered as a relatively good selective V 2 agonist (Table 11).
Selective vasopressin V 1a receptor agonists (Table 4) In rat bioassays, [Phe 2 ]OVT (peptide 2; Table 4) is a fairly potent vasopressor agonist (81). Its vasopressor (P) activity is 124 units ⁄ mg. In antidiuretic (A) assays, it exhibits only 0.55 units ⁄ mg. Its P ⁄ A ratio is 225 (81). Thus, for many years, it has been considered to be a selective V 1a agonist and has been widely used as a selective V 1a agonist. However, based on its rat V 1a receptor affinity   data in Table 11, it is not selective for the rat V 1a receptor in this assay. In this regard, the selective V 1a agonist F-180 (82), which is a highly selective V 1a agonist in rat bioassays (peptide 3; Table 4), is even more puzzling. In rat receptor assays (  (Table  11). In the rat, this agonist is very specific for the V 1a receptor compared to the V 2 receptor (selectivity higher than 800), yet it has not been tested for the OT and V 1b receptors. This intriguing new V 1a agonist is not yet available to other scientists for use as a pharmacological research tool.
V 1b receptor agonists (Table 5) AVP was synthesised in 1954 (3) (  Table 5 lists four analogues of dAVP (peptides 1-4) that exhibit high affinities for both the rat and human V 1b receptors. d[Cha 4 ]AVP (peptide 1) was the first V 1b agonist that was shown to be selective for the human V 1b receptor (69). d[Leu 4 ]AVP (peptide 3) has later been shown to be a selective agonist for the human V 1b receptor (72). Both (peptides 1 and 3) exhibit high affinities for the rat V 1b receptor. However, they also possess high in vivo antidiuretic activity. Thus, neither is a selective V 1b agonist in the rat.
Replacement of the Arg 8 residue in (peptides 1 and 3) by a Lys 8 residue to give d[Cha 4 ,Lys 8 ]VP (peptide 2) and d[Leu 4 ,Lys 8 ]VP (peptide 4), respectively, resulted in the first peptides that are selective V 1b agonists in the rat (73,74). It was subsequently shown that both peptides 1 and 2 are also highly selective for human V 1b receptors (86). It bears noting that d[Leu 4 ,Lys 8 ]VP had been reported to be a weak antidiuretic V 2 agonist ⁄ weak vasopressor agonist in the rat (87), 30 years before the V 1b receptor was first predicted and ⁄ or cloned (88 (1,20,62,73,89,90). Furthermore, d[Leu 4 ,Lys 8 ]VP has been utilised in the design of a series of fluorescent ligands for the V 1b receptor (47).  ]AVP (peptide 2; Table 6) and d(CH 2 ) 5 [Tyr(Me) 2 ,Dab 5 ]AVP (peptide 3; Table 6), respectively. Both peptides are devoid of anti OT activity in vivo (92). Although both peptides are much less potent than d(CH 2 ) 5 [Tyr(Me) 2 ]AVP as V 1a antagonists, because they lack anti OT potency in vivo, they are highly selective for V 1a receptors in the rat. Their use is recommended for in vivo studies that require discrimination between V 1a and OT receptors in the rat.

Nonselective and selective cyclic and linear V 2 ⁄ V 1a antagonists for rat receptors
It was not until 1981, almost 30 years after the first laboratory synthesis of OT, that the first cyclic AVP V 2 ⁄ V 1a antagonists were reported (8) ( Table 7). Six years later, the unexpected discovery of the first linear V 2 ⁄ V 1a antagonists was reported (93). The early cyclic and linear V 2 ⁄ V 1a antagonists were nonselective for V 2 receptors. Further modifications of the early cyclic V 2 ⁄ V 1a antagonists led to the discovery of selective cyclic V 2 antagonists (94). Some of the In vitro pA 2 values represent the negative logarithm to the base 10 of the average molar concentration [M] of the antagonist that reduces the response to 2 · units of agonist to the equal the response seen with 1 · units of agonist administered in the absence of the antagonist. b In vivo pA 2 values are estimated because the molar concentration for the in vivo pA 2 is estimated by dividing the effective dose (ED) by the estimated volume of distribution of (67 ml ⁄ kg) (52). ED is defined as the dose (nmol ⁄ kg intravenously) of the antagonist that reduces the response to 2 · units of agonist to the response with 1 · units of agonist administered in the absence of the antagonist. c ND, not detectable (weak agonist, < 0.03 U ⁄ mg).  Aaa, adamantaneacetyl; Eda, ethylenediamine; 4-HO-Phaa, 4-hydroxyphenylacetyl. *In vivo anti-oxytocin (OT) potencies were reported previously (10), a most commonly used nonselective cyclic and linear V 2 ⁄ V 1 antagonists (peptides 1, 2, 7) and selective cyclic V 2 antagonists (peptides 3-6) are given in Table 7. Peptide 6 (101) has been very useful for the design of a lanthanide cryplate-labelled ligand as a fluorescent probe for measuring receptor dimerisation (48) Peptide 8 (HO-LVA) (95), a potent linear V 1a antagonist, has served as a precursor for the radioiodinated V 1a ligand [ 125 I]HO-LVA (97). This radioligand has found widespread use as a selective probe for V 1a receptors. Its affinities for rat receptors are given in Table 12. A number of these V 2 ⁄ V 1a antagonists (peptides 1-3; Table 7) exhibit oxytocic antagonism in vivo. The remaining peptides 4-8 have not been evaluated in anti-OT assays. Caution should be exercised in using any of the peptides in Table 7 as selective V 2 ligands. The affinities of peptides 1 and 5 for the human and rat V 2 receptors are given in Table 12.
Some nonselective and selective oxytocin antagonists (Table 8) The OT antagonists listed in Table 8 have all found widespread use as pharmacological tools. Under the tradename Tractocile, atosiban (peptide 1) has been approved for clinical use in Europe for the prevention of premature labour (102). All of these OT antagonists exhibit varying degrees of anti-V 1a potency in the rat. Thus, they are far from being selective. Indeed, d(CH 2 ) 5 [Tyr(Me) 2 ]OVT, one of our original OT antagonists (103), is five-fold more potent as a vasopressor antagonist than as an OT antagonist (103). Peptide 5, desGly-NH 2 , d(CH 2 ) 5 [D-Tyr 2 ,Thr 4 ]OVT (104) is the most selective OT antagonist in Table 8. It has been used in a variety of studies. These are listed under 'Research Uses'. This OT receptor antagonist also exhibits a high affinity and selectivity for the human OT receptor (  2 ,Thr 4 ,Tyr-NH 2 9 ]OVT (105) has found widespread use as a probe of OT receptors.
Nonpeptide oxytocin antagonists and a nonpeptide oxytocin agonist (Table 10) A number of companies have been active in this area. In the mid-1990s, Merck reported a number of promising nonpeptide OT antagonists (46,130) (Table 10). Most notable were L-368,899 (No. 1; Table 10) and L-371, 257 (No. 2; Table 10). Both of these failed in clinical development for the treatment of premature labour. Merck subsequently abandoned its nonpeptide OT antagonist programme.  The use of radiolabelled molecules, agonists and antagonists for characterising receptor affinities for OT and VP (Tables 11 and 12)

Some history
Initially in the 1960, the affinity, selectivity and potency of analogues for the different VP ⁄ OT receptors were deduced by in vivo bioassays such as oxytocic, antidiuretic and pressor tests (see above), which reflected their activity through the OT receptor, V 2 and V 1a receptors, respectively. The characterisations performed at that time did not take into account the V 1b receptor, which was discovered only in the 1980s (142). As noted above, this led to a long delay in the discovery of selective ligands for the V 1b receptor. Nevertheless, the use of bioassays allowed the identification of key structure ⁄ function relationships of a large number of analogues and still represents a milestone in our understanding of OT ⁄ AVP selectivity (1,(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18). Furthermore, because the data obtained using these in vivo tests integrated several ADME (Absorption, Distribution, Metabolism and Elimination) parameters, the values obtained reflect the in vivo physiological activity of the peptide being studied. These values sometimes differed from those obtained by in vitro pharmacological tests.
In the 1970s, the development of radiolabelled VP ⁄ OT analogues (143) and the discovery of second messenger cascades such as cAMP (67), calcium and inositol phosphate (144) made possible the determination of more reliable pharmacological parameters reflecting more precisely the interaction between analogues and their specific receptors. Binding assays with radiolabelled ligands conducted on plasma membrane preparations allowed the determination of the affinity (K d ) of a given molecule for a given VP ⁄ OT receptor subtype, a parameter which intrinsiquely characterises the analogue ⁄ receptor association (145). Second messenger measurements allowed the characterisation of its functional activity in order to measure precise functional effects. Comparison of the affinities of one analogue for the all receptors of the VP ⁄ OT family allowed the determination of its selectivity towards a given receptor iso- form. Moreover, the ability of a given analogue to activate, inhibit or leave unaffected second messenger production in cell cultures, provided important insights into its pharmacological status (agonist, partial agonist, pure antagonist, inverse agonist). Such classical pharmacological assays have efficiently served the scientific community for the last three decades and allowed the characterisation of the numerous VP ⁄ OT analogues designed and synthesised during this period (1).

Conundrums posed by pharmacological data
From all the in vitro and in vivo pharmacological studies carried out on OT and VP agonists and antagonists, three intriguing fea-tures have emerged; namely: (i) lack of receptor selectivity; (ii) species differences; and (iii) in vitro in vivo difference. In Tables 11  and 12, we have listed the most commonly used agonists and antagonists available for each VP ⁄ OT receptor isoform in three mammalian species: human, rat and mouse. The affinities (K i ) of these isoforms have been measured using classical pharmacological tests and their agonist or antagonists properties determined using classical second messenger assays. As proposed in previous reviews (156,157), receptor subtype selectivity can be defined, within a single species, on the basis of the ability of a compound to bind to a single VP ⁄ OT receptor isoform with a nanomolar affinity, at the same time displaying, for the three other receptor isoforms, an affinity at least two orders of magnitude Affinities values (in nM) were obtained on cells expressing the oxytocin receptor (OTR), V 1a , V 1b and V 2 receptor subtypes. Light grey boxes report affinities of selective ligands for a particular receptor subtype as defined by the following selectivity criteria: to be selective a ligand must have a K or K d for that receptor subtypes two orders of magnitude lower than for the other three receptor subtypes in the same species, b Rounded off or representative values from all the data available.
lower. The compounds that fullfill these requirements are highlighted with light grey in Tables 11 and 12. It immediately appears from this criterion that only a very few analogues are selective. A major problem is also that selectivity is not conserved among species as a result of subtle but nevertheless crucial differences in receptor pharmacology. Despite these limitations, the use of selective compounds still represents the best experimental strategy to unambiguously characterise VP ⁄ OT receptors in a given biological sample, keeping in mind that receptor selectivity for any given compound is: (i) strictly dependent upon the receptor species considered; (ii) usually lost if high doses (100-fold the K i ) of a selective compound are used; and (iii) dependent upon the biological models tested. Experiments performed on membrane preparations or on cell cultures generally need lower concentrations of selective analogue compared to experiments performed on organ slices, where drugs need to diffuse within the tissues and may be rapidly degraded. It should also be noted that the pharmacological profile of any given compound determined by classical tests on membranes or cell models cannot be directly translated in vivo without adequate controls. Adsorption, distribution in different biological compartments, and metabolism greatly interfere with the biological activity of drugs, sometimes completely altering their pharmacological properties.
For example, [Phe 2 Orn 8 ]VT (also known as [Phe 2 ]OVT), which does not display any V 1a selectivity in classical binding experiments (Table 11), has been characterised as a selective V 1a agonist in vivo in rats (Table 4) (81).
Concerning selective agonists, it should also be noted that a major difference exists between the two natural hormones, OT and VP. Although OT is selective for the human OT receptor, VP is not, because it binds with similar affinities to V 1a , V 1b V 2 and OT receptors. This may explain why VP may trigger physiological functions in vivo via OT receptors, as described previously (158). However, fully characterised selective agonists for human V 1a receptor (F 180, FE202158), human V 1b receptor (d[Cha 4 ]AVP); rat V 1b receptor (d[Leu 4 ,Lys 8 ]AVP), rat OT receptor [Thr 4 ,Gly 7 ]OT and rat V 2 receptor (dDAVP) are now available (Table 11). For the rat, V 2 receptor d[Thi 3 ]VDAVP (15,159) appears to be the best selective agonist as a result of its good V 2 versus V 1b selectivity.
Among the several antagonists reported and currently employed, only a few have been fully characterised and have been demonstrated to be selective whithin a species (Table 12).
Among the OT receptor antagonists, SSR126768A has been shown to be a very selective antagonist for both human and rat OTR and GSK 221149A for human OT receptor. Manning compound is relatively selective for the rat V 1a receptor (but not for the human V 1a receptor) for which SSR49059 should be preferred. Finally, SSR149415 (147) is selective for both the human and the rat V 1b receptor isoforms, whereas SSR121463(A) is highly selective for the human V 2 receptor. Concerning the SSR149415 and SSR49059, it should be noted that different laboratories have obtained different values for their affinities, probably depending on the binding assay employed (i.e. competition against a radioactive agonist or antagonist) (89). In our opinion, the values obtained in competition experiments using radiolabelled agonists will better correlate with biological antagonistic activity in vitro and in vivo and should be preferred.
Until now, other analogues commonly employed as 'selective' have not been fully characterised and, when they are used at high doses, could lead to ambiguous results in species in which their pharmacological properties have not been assessed. It should be noted that the pharmacology of OT ⁄ VP analogues on mouse receptors is still very limited, representing a gap that needs to be filled; in particular, for the relevance that genetically-modified mouse models have acquired in translational medicine.
The lack of selective analogues for some rat, mouse and human receptor isoforms makes the design and synthesis of new molecules very necessary. The restriction of radioactivity approaches in laboratory practice and the need to easily test a large number of molecules led to the development of new assays using the gene reporters. Such tests using the measurement of reporter gene activities allows an easy screening of a large number of molecules and rapid identification of 'lead molecules' (83).
Yet, these 'in vitro methodologies' also have some limitations. First, they can be used only in transfected cells and not in native models. Moreover, according to the second messenger cascade associated with the receptor being considered (cAMP for the V 2 , InsP3 for the V 1a , V 1b and OT receptor isoforms), such assays require the use of different reporter genes. This may introduce a bias in the determination of receptor selectively. It is also well known that assays using luciferase gene expression and luciferase activity measurements involve a strong amplification of the initial receptor-mediated second messenger accumulation. This prevents the good determination of the agonist or partial agonist properties of the analogue tested. One needs to be aware of the limitations of these recent 'in vitro methodologies' and to verify, using classical pharmacological tests, the selectivity, affinity and functional potencies of the lead compounds characterised by this approach. Obviously, to move to clinical development, the best approach would be to test the compounds of interest by in vivo technologies similar to those used to evaluate virtually all of the peptides in Tables 1, 3-8.

New technologies for screening more selective VP ⁄ OT analogues
Recently, new physical techniques involving label-free biosensors have been proposed for pharmacological screening of muscarinic and corticotrophic analogues (165). These methods are based on the measurement of cell shape changes induced by ligand-receptor interactions. Such techniques have the advantage of being performed on native cells and do not require the use of radioactive molecules. Their efficiency for testing new VP ⁄ OT molecules may represent another alternative for screening new analogues.
Finally, a new bioluminescence or fluorescence resonance energy transfer (BRET or FRET) approach in which analogues could be screened for their capability to promote receptor coupling and activation of single G-protein isoforms has been recently applied to the human OT receptor (166). This technique allows the precise characterisation of which G protein is associated with which recep-tor isoform. Thus, for example, d(CH 2 ) 5 [D-2-Nal 2 ,Thr 4 ,Tyr-NH 2 9 ]OVT (OTA) and atosiban (160) (Table 12) were found to be entirely biased respectively toward Gi1 or Gi3 activation (166). However, this technique cannot be used on native tissues or primary cultures.
The recent development of fluorescent ligands for a better knowledge of central and peripheral VP ⁄ OT receptors

Design and use of classical fluorophores
Receptors of the AVP and OT family are important in the regulation of the stress processes (167). Centrally, the V 1a , V 1b and OT receptors have been involved in stress and especially in learning and memory processes. Important data have been obtained by the use of knockout animals but, after a period of cloning and pharmacological characterisation in the last decade, it became necessary to elucidate the distribution of these receptors to better understand their central functions in vivo.
Although several publications describe the AVP V 1a , V 2 and OT receptor distribution by using autoradiography (105,168,169) or immunodetection (170), the lack of selective V 1b radio-labelled VP analogues or of receptor antibodies has hindered progress in the detection of receptor distribution in native tissues. Results obtained by molecular approaches such as reverse transcriptase-polymerase chain reaction (62,88) or mRNA detection by in situ hybridisation (171)(172)(173), although more accurate, did not provide clear information regarding the brain regions detected by immunostaining. Thus, developing fluorescent ligands to decipher AVP receptor distribution in the brain and at the periphery, and to study molecular interactions such as receptor dimerisation, appeared as an absolute necessity.
Various fluorescent analogues of AVP and OT have been synthesised for several receptors of the VP ⁄ OT family (174). Thus, good fluorescent V 1a and OT ligands have been produced, although no good fluorescent specific ligand was available to selectively detect central and peripheral V 1b receptors.
In our previous work (69,73,74), by replacing the glutamine 4 of the natural AVP with a cyclohexylalanine or a leucine, the arginine 8 by a lysine and by removing the NH 2 of the cysteine 1 to increase stability towards aminopeptidases, we produced analogues showing an increased selectivity for the V 1b receptors (Table 5). d[Leu 4 ,Lys 8 ]VP (Peptide 4; Table 5) was found to be selective for the rat V 1b receptors and, to a lesser extent, for the human hV 1b receptors, conserving a nanomolar affinity for these receptor isoforms (73,74). We have taken advantage of the Lys 8 residue in d[Leu 4 ,Lys 8 ]VP with its epsilon NH 2 group to introduce fluorophores on its side chain. This allowed us to create fluorescent tools that would conserve the pharmacology of the d[Leu 4 ,Lys 8 ]VP, to resist degradation and to selectively decorate the plasma membrane of Chinese hamster ovary cells expressing V 1b and ⁄ or OT receptors with an excellent resolution (47).
Different fluorophores were attached to the d[Leu 4 ,Lys 8 ]VP: First, the antraniloyl group (Atn), a small fluorescent molecule of 97 Da highly sensitive to microenvironmental changes (175) may also be a good donor in FRET experiments to identify V 1b receptor homodi-mers in vivo. We have also selected the Alexas (Molecular Probes) for their brightness and their resistance to photobleaching (176). We have used Alexa 488 and Alexa 647, with the latter being one of the brightest fluorescent molecules reported so far (177).
The pharmacological properties (binding, coupling to phospholipase C) of fluorescent analogues of d[Leu 4 ,Lys 8 ]VP indicate that they conserved a very good selectivity for V 1b versus V 1a and V 2 receptors, and remained full agonists. These properties allow receptor labelling and measurement of biological activity at the cellular level. Thus, these new fluorescent analogues are promising tools for the detection of functional V 1b or OT receptors in human (47) and in rat native tissues.

Use of long life fluorophores
However, it should be noted that, except for very recent ones, all the ligands previously reported were designed with classical fluorophores, exhibiting short-lived fluorescence properties (fluorescence half-time life in the 10 ns range). Most of them were essentially used to follow internalisation in cell lines (146,178) or to label receptors in a native context (179). Interestingly, a first nonpeptide antagonist with a nanomolar affinity for the human V 2 receptor has been developed (180). This ligand will find application in fluorescence polarisation-based binding assays aiming to screen for small organic molecule libraries.
Recently, fluorescence-based strategies have been extensively used to study molecular interactions. Thus, the FRET approach was used to demonstrate G protein-coupled receptors oligomerisation (181,182). Regarding AVP and OT receptors, various experimental approaches based on chimeric receptor expression in cell lines have been developed to analyse receptor oligomers. The studies have used receptors fused either to small tags recognised by fluorescent antibodies (183,184), or to bioluminescent or fluorescent proteins (185,186), or to suicide enzymes (48). However, these strategies were not relevant for proving the existence of such receptor complexes in native tissues. Therefore, a FRET strategy based on the indirect labelling of receptors with fluorescent donor and acceptor ligands has been recently developed (48). Unfortunately, because of the overlap between excitation and emission spectra of the donor and acceptor fluorophores on one hand, and of the high autofluorescence of the biological preparation on the other hand, specific FRET could hardly be detected. To improve the signal-to-noise ratio, lanthanide cryptate-labelled ligands were designed and characterised. Despite the size of the cage, these ligands still display very good affinities for the V 1a and OT receptors (48). Lanthanide cryptates display interesting fluorescent properties because they have a fluorescent half-time life of approximately 1 ms (i.e. 100 000-fold greater than classical fluorophores), allowing time-resolved FRET experiments to be set up (187). Experiments using these new probes have been performed not only on AVP V 1a and V 2 receptors and on OT receptors expressed in cell lines, but also on OT receptors naturally expressed in lactating rat mammary gland. The sensitivity is such that it has been possible to prove the existence of OT receptor dimers in this latter native tissue (48).
These newly-synthesised ligands and those that exhibit high quantum yield have also been used to develop original binding assays. These assays, based on time-resolved FRET between compatible fluorophores carried by tagged receptors and ligands, display very good sensitivities and are safer than radioactive-based assays (188)(189)(190).
Therapeutics uses of peptide and nonpeptide oxytocin and vasopressin agonists and antagonists (Table 13) Table 13 lists the seven peptides and two nonpeptide drugs in the OT and AVP field that have been approved for therapeutic use. Numerous recent studies point to the use of OT as a potential new therapy for the treatment of a broad range of psychiatric disorders (24,33,34). Also, a very recent report suggests the exciting prospect that OT may have potential for the treatment of human obesity and type 2 diabetes (191).
To date, only two nonpeptides, the VP V 2 ⁄ V 1a antagonist conivaptan and the VP V 2 antagonist tolvaptan have been approved for clinical use (112-115, 123, 192). The nonpeptide OT antagonist retosiban (132,133) is currently in a Phase II clinical trial. Clinical trials with other nonpeptide VP, V 1a and V 1b antagonists shown in Table 9 and with the other OT antagonists and the OT agonist shown in Table 10 have been terminated (44). No new nonpeptides in this field are currently in clinical trial. Thus, the early promise of nonpeptides as therapeutic agents in this field (45,46) has clearly not been realised. This should be a cautionary tale for those in pharmaceutical companies (138), in granting agencies (139) and on study sections (44) who have strongly promoted the development of nonpeptides over peptides as therapeutic agents. This philosophy led to the abandonment of peptides as potential drugs by Big Pharma almost two decades ago. Clearly, it is now time for Big Pharma to reassess the value of peptides as therapeutic drugs (44). In this regard, recent progress in the development of peptide drugs (193,194) provides very compelling support for the thesis that peptides are clearly superior to nonpeptides as therapeutic agents, thus bolstering the case for continued support for the design and syn-thesis of peptides as potential therapies. In this regard, in the OT field, there is an urgent need for functionally selective OT ligands (166) and for a long lasting OT agonist as a potential therapy for the treatment of autism and other anxiety disorders (139).

Research uses of peptides
During the period 1980-2012, over 3000 samples of OT and AVP agonists and antagonists from the Manning laboratory have been and continue to be donated as research tools to over 700 investigators (some multiple times) in the USA and worldwide for their own independent studies. Studies carried out with these donated peptides and with those purchased from commercial suppliers, such as Sigma, Bachem and Peninsula, have resulted in more than 2000 publications by these and other investigators during this period. Examples of studies carried out since 2008 with some of these peptides are available (90,117,.

Research uses of nonpeptides
With the exception of the Sanofi nonpeptide V 2 antagonist satavaptan, all of the nonpeptide AVP antagonists listed in Table 9 are now available from Tocris or other suppliers. To date, a small number of studies that have utilised the Sanofi V 1a antagonist relcovaptan, the Sanofi V 1b antagonist nelivaptan and the Sanofi V 2 antagonist satavaptan have been reported (117-120, 125, 126) (Table 9).
The two Merck nonpeptide OT antagonists L-368,899 and L-371,257 and the Pfizer nonpeptide OT agonist Way-267464 shown in Table 10 are all available from Tocris. The Glaxo SmithKline nonpeptide OT antagonist GSK-2211149A (Retosiban) is also available from a number of suppliers. The Pfizer nonpeptide OT antagonist WAY-162720 is not yet commercially available. A number of studies that have utilised some of these nonpeptide ligands as research tools have been reported (138)(139)(140)(141)145) (Table 10).

Conclusions
In our 2008 review (1), we examined the status (as of 2007) of both peptide and nonpeptide agonists and antagonists of the OT receptor and of the VP V 1a , V 1b and V 2 receptors as: (i) research tools and (ii) therapeutic agents. Although the research uses of both peptide and nonpeptide ligands have continued to grow during the intervening 4 years, by contrast, the therapeutic development of nonpeptide AVP and OT antagonists has been drastically curtailed. Merck, Pfizer and Sanofi have all abandoned their nonpeptide programmes. The nonpeptide V 2 ⁄ V 1a the antagonist, conivaptan and the nonpeptide V 2 antagonist tolvaptan, which have been approved by the Food and Drug Administration, have not as yet found widespread acceptance in the clinic (113,192). Pfizer has also abandoned its nonpeptide OT agonist programme (44). It remains to be seen how the Glaxo SmithKline nonpeptide OT antagonist retosiban will fare in its current Phase II clinical trial. All in all, since our 2008 review (1), interest in the development of nonpeptides as therapeutics has greatly diminished. On the other hand, as noted above, Ferring has a promising V 1a agonist (Table 4) and a promising OT agonist (Table 1), awaiting clinical development. The design and synthesis of: (i) functionally selective OT peptides and (ii) of a long lasting OT analog as a potential therapy for autism spectrum disorders and other anxiety disorders remain as pressing needs in this field. Both OT and VP peptides and nonpeptides are continuing to be very valuable research tools. In this regard, we have addressed here the issues of the: (i) lack of receptor selectivity, (ii) species differences and (iii) in vitro-in vivo differences, all of which need to be taken into account when using a given peptide or nonpeptide ligand as a research tool. Finally, the development of new fluorescent ligands as powerful new tools for localising and characterising OT and VP receptors, which we have presented here, points to the continued usefulness of OT and AVP peptide ligands (agonists, antagonists and fluorescent derivatives) as incisive molecular pharmacological probes.