RFamide Peptides: Structure, Function, Mechanisms and Pharmaceutical Potential

Neuropeptides characterized by a common carboxy-terminal arginine (R) and an amidated phenylalanine (F) motif (designated RFamide peptides) were originally discovered in invertebrates [ 1 ]. In mammals it is now known that there exist at least five genes, encoding the family members and the five G-protein coupled receptors through which RFamide peptides act. All these peptides show a great diversity, regarding their N-termini and hence a wide pattern of biological activity. The initially identified tetrapeptide FMRFamide was isolated from the ganglia of the clam as a cardioexcitatory peptide [ 1 ]. Significant improvements in the molecular characterization of the RFamide peptide system have been made since then. Nowadays the rising number of RFamide peptides found in mammals can be subdivided into five groups: the neuropeptide FF (NPFF) group, the gonadotropin-inhibitory (GnIH) group, the 26RFa group, the kisspeptin/metastin group and the prolactin-releasing peptide (PrRP) group. By applying a reverse pharmacological method and by searching of DNA sequence databases several other structurally similar peptides have also been identified ( Table 1 ). The first reported vertebrate member was LPLRFamide [ 2 ], and the first mammalian RFamide peptides were neuropeptide FF and neuropeptide AF [ 3 ]. Both were purified from bovine brain extracts [ 3 ], showed several biological activities in vivo in mammals [ 4 ] and are encoded by a single gene [ 5 ]. GnIH homologs were isolated from extracts of the human hypothalamus by immunoaffinity purification, and furthermore identified as human RFamide-related peptide 1 (RFRP-1) and human RFRP-3 by mass spectrometry [ 6 ]. Both, 26RFa and the longer form of 26RFa, termed 43RFa (QRFP), have been isolated from the human hypothalamus and spinal cord [ 7 ]. PrRP was identified in 1998 by a reverse pharmacology approach [ 8 ]. Up to now two equipotent isoforms of different N-terminal length are known, PrRP31 and PrRP20 [ 9 ]. The naturally occurring kiss RFamide peptides, consisting of 54-, 14-, or 13-amino acids were isolated from human placenta [ 10 ], whereas the highly potent kisspeptin-10 was synthesized [ 14 ]. Due to the prediction that all three endogenous peptides derive from KiSS-1 [ 11 ], a metastasis suppressor gene for melanoma cells, they were designated as kisspeptins. It should be noted that Hori et al. [ 12 ] named the same peptide metastin, which derives from its initially discovered function to suppress metastasis, but for clarity kisspeptin is used preferentially in this review.

Several orphan receptors turned out to respond to different members of the RFamide peptide family ( Table 2 ). The PrRP receptor was the first putative RFamide receptor that was identified and initially described as GPR10 [ 8 ]. In 2000, the putative receptors for NPFF and NPAF were identified and named NPFF1/GPR147 and NPFF2/GPR74 [ 13 , 16 , 17 ]. Independently, two groups could show that both, 26RFa and 43RFa, act as natural ligands of the orphan receptor GPR103 [ 18 , 19 ]. The kisspeptin receptor or KiSS1-derived peptide receptor was formerly discovered as orphan G-proteincoupled receptor GPR54 [ 20 , 22 ]. For the explanations that follow we refer to the protein names of the corresponding receptors ( Table 2 ). So far there are only few mutagenesis data available. We decide to focus on one residue on the top of transmembrane helix (TMH) 6, which is known to be present and/or important for receptor activation in several related receptor families (e.g. the NPY receptor family). Because all ligands share an amidated dipeptide motif at their C-termini with the positively charged arginine side chain, an acidic residue has been proposed to act as a counterpart. Therefore, first investigations were made on an acidic residue 6.59 on the top of TMH 6, according to the nomenclature of Ballesteros and Weinstein [ 23 ]. It was recently shown by immunohistochemistry that cells expressing the human NPFF 1 R are localized in the human hypothalamus and surrounding areas [ 24 ]. It is now clear, that most human forebrain nuclei contain such cells, supporting the idea that RFamide peptides play a role in central coordination of neuroendocrine and autonomic responses in humans. The NPFF 1 R commonly activates the G i/o signal transduction pathway [ 13 , 16 ]. The human receptor was isolated from a human spinal cord cDNA library and analysis by sequencing led to a coding sequence of 1290 bp, and a predicted protein with a length of 430 amino acids [ 16 ]. Comparing the primary sequence of human NPFF 1 R with other known GPCRs showed, that NPFF 1 R is most similar to human orexin1 (37% identity), human orexin2 (35%), human neuropeptide Y receptor subtype 2 (NPY 2 R) (34%), human chole-cystokinin A (CCKA) (34%), human NPY 1 R (32%), mouse GIR (32%), human prolactin-releasing hormone receptor (32%), and human NPY 4 R (31%) [ 16 ]. RFRP-1 and RFRP-3 efficiently inhibit production of forskolin-induced cAMP accumulation in CHO cells, expressing the human NPFF 1 receptor [ 13 ]. Furthermore, pharmacological characterization of hNPFF 1 -transfected SH-SY5Y cells was performed by measuring the intracellular cAMP content [ 25 ]. By using a [ 35 S]GTPγS binding assay, the functional activities of RFamide peptides were investigated in CHO cells stably expressing the NPFF 1 receptor [ 26 ]. To the best of our knowledge, there is just one receptor mutant published so far. In studies regarding the related NPY receptor family, Merten et al. could clearly demonstrate that the acidic residue 6.59 on the top of TMH 6 plays an important role in ligand binding [ 27 ]. Thus, Findeisen et al. presumed a strong ionic interaction to be involved in ligand recognition and receptor activation at position 6.59. When the derived Asp 6.59 mutants were investigated [ 28 ], the introduction of an elongated charge conserving glutamate revealed a minor loss in receptor activation, whereas the replacement of Asp with Ala or Asn led to a significant loss in affinity and receptor activation. The need for a negatively charged side chain at position 6.59 is underlined by the introduction of Arg, resulting in strongly decreased affinities and a not fully activatable hNPFF 1 R mutant. Another part of the binding pocket is proposed to be built up by residues located in the upper sections of the TMHs. Further investigations revealed evidence that other hydrophobic residues are not involved in a direct interaction. By mutating putative residues of impact to alanine, we merely found a minor loss of receptor activation for Phe 5.43 , Phe 7.37 , Phe 7.44 (unpublished data). By autoradiography and by the use of a radiolabelled NPFF analog, [ 125 I](1DMe)Y8Famide, it could be shown that NPFF receptors are localized in the central nervous system [ 29 , 30 ]. The NPFF 2 R mostly activates the G i/o signal transduction pathway [ 16 , 17 ]. The human full-length receptor was amplified from spinal cord cDNA and further analysis by sequencing led to a coding sequence of 1260 bp, and a predicted protein with a length of 420 amino acids [ 16 ]. Both NPFF receptors share 50% amino acid identity. Pharmacological characterization of human NPFF 2 receptor was performed by measuring the cAMP levels in CHO cells as well as in HEK293 cells transfected with hNPFF 2 R [ 17 ]. Furthermore, the anti-opioid effects on Ca 2+ mobilization were investigated by measuring Ca 2+ contents in SH-SY5Y cells transfected with the human NPFF 2 R [ 31 ]. Additionally, the [ 35 S]GTPγS binding was studied in CHO cells stably expressing the human NPFF 2 R [ 26 ]. To compare the receptor subtype specific activation the same acidic residue on the top of TMH 6 was investigated by Findeisen et al. [ 28 ]. The introduction of glutamate at position 6.59 revealed an equipotent minor loss in NPFF 2 R activation as for NPFF 1 R. Regarding NPFF receptor subtype selectivity the modifications of Asp to Asn and Ala was better tolerated at the hNPFF 1 R, compared to the hNPFF 2 R. This clearly indicates that the hNPFF 2 R is more sensitive at this receptor site. The replacement of Asp 6.59 with Arg resulted in a not fully activatable hNPFF 2 R variant, which demonstrates the necessity of a negatively charged amino acid. Further investigations regarding hydrophobic residues in the upper transmembrane regions revealed no involvement in receptor activation. By mutating residues of interest to alanine we could see no loss of receptor activation at position Phe 5.43 or Phe 7.40 (unpublished data). The QRFP receptor is an orphan G-protein coupled receptor and was originally cloned from a human brain cDNA library. It has been detected widely throughout the brain, heart, kidney, retina and testis [ 19 ]. GPR103 encoded a 455 amino acid protein and phylogenetic analysis showed that the orphan receptor shares identities with peptide-binding receptors, including NPFF 2 R, NPY 2 R and galanin GalR1 receptors (34–38% in the TM regions). The QRFPR commonly activates the G q /G i signal transduction pathway. With regards to the conserved acidic residue at position 6.59, a glutamate is present instead of an aspartate. Nevertheless, by the replacement with alanine, we could see the same loss (∼60-fold) in activity as for the NPFF 1 R (unpublished data). Replacement of a hydrophobic residue (Phe 5.41 ) by alanine revealed no involvement in receptor activation since no decreased EC 50 -value was detected (unpublished data). The kisspeptin receptor was originally cloned from rat brain, sharing 45% identity with the galanine receptors [ 22 ]. In humans, reverse transcriptase polymerase chain reaction has revealed, that GPR54 was highly expressed in placenta, pituitary, pancreas, and spinal cord, suggesting an important role in the regulation of endocrine function [ 10 , 15 ]. Furthermore, by immunohistochemistry kisspeptin receptor protein expression was detected in the cerebral cortex, thalamus, pons-medulla and cerebellum [ 20 ]. Together with the use of radioligand binding Mead et al. found expression in aorta, coronary artery and umbilical vein [ 32 ]. Functional assays were performed in CHO cells transfected with the kisspeptin receptor [ 12 ] and in vitro pharmacology was carried out by using human isolated vessels [ 32 ]. Up to now it is known that the activated kisspeptin receptor acts via G q /G 11 signal transduction pathways, leading to the stimulation of phospholipase C and consequently to calcium mobilization. Additionally, activation of the kisspeptin receptor has been shown to stimulate arachidonic acid release [ 10 ]. While the acidic residue 6.59 is conserved throughout the whole RFamide peptide receptor family, in the human kisspeptin receptor an alanine is present at that position. To our knowledge, no mutational studies have been performed so far. It should be noted that by using polymorphism scanning and typing of KiSS-1/GPR54 gene, Chu et al. recently unveiled a couple of harmful polymorphisms [ 33 ]. This includes the mutation of residue Phe in position 272 to Ser in the GPR54 gene, a loss-of-function mutant, which is associated with familial normosmic IHH (idiopathic hypogonadotropic hypogonadism) [ 34 ]. Underdeveloped external genitalia and impuberism point to the major role of GPR54 in the activation of the gonadotropic axis from intrauterine life to adulthood [ 34 ]. The prolactin-releasing peptide receptor was originally isolated from rat hypothalamus [ 35 ] but has been detected widely throughout the human and rat brain [ 36 ]. After binding of the endogenous ligand the activated receptor most commonly activates the G q signal transduction pathway [ 9 ]. The only manuscript concerning mutagenesis studies which is published so far reports that the PrRP receptor has a PDZ domain binding motif in its C-terminal tail (-SVVI) [ 37 ]. In a mutational study with the prolactin releasing peptide receptor we were able to show the involvement of the aspartic residue at position 6.59 (unpublished data). By replacement with alanine, a 22-fold loss of receptor activity upon stimulation with PrRP was observed. As mentioned above a second part of the binding pocket is proposed to be built up by residues, located in the upper sections of the transmembrane helices. Therefore, in a first study we investigated the involvement of hydrophobic residues. We could not detect a decreased EC 50 -value, when Phe 2.72 , Trp 5.40 , Phe 7.31 or Trp 7.40 was mutated to alanine upon stimulation, which revealed no involvement in ligand binding and receptor activation (unpublished data).

In recent years, the essential role of RFamide peptides became more and more clear as the number of physiological functions and effects, the peptides participate in, were growing. The regulatory pathways of the RFamide peptides are involved in prominent nodal points like the cardiovascular system, feeding behavior, locomotion, nociception, energy homeostasis reproduction or cancer metastasis. On the one hand, perturbation within these signaling pathways might lead to serious syndromes but on the other hand the RFamide family opens the field to modulate a plethora of undesired effects in a positive manner. This issue is becoming increasingly important, because cancer, eating disorders, failed reproduction or the cardiovascular diseases are problems of our modern society and an ascending role will be their treatment by drugs. To address the correct signaling pathway in a desired manner it is indispensable to further elucidate the role of the distinct RFamide peptides. Therefore, highly selective agonistic and antagonistic compounds/peptides are essential and still needed, as the crosstalks within this family, but also to the NPY system are quite complex and not fully understood. First steps were made for NPFF and kisspeptins, but this path has to be followed further to benefit from the huge modulatory potential of the RFamide peptide system. Herein, we summarize the diverse physiological roles of the RFamide peptide family and provide detailed insight into the latest structure-activity and structure-affinity relationship studies.

NPFF was initially identified to attenuate morphine induced antinociception and to evoke an increased sensitivity to pain, called hyperalgesia [ 3 , 38 ]. Since then, the range of activities has been expanded and is associated with a plethora of physiological and behavioral phenomena. In rats, NPFF was shown to participate in the modulation of the cardiovascular system, basically inducing increased blood pressure, bradycardia or inhibition of the cardiac component of the baroreceptor reflex [ 39 - 45 ]. Further studies revealed an anorectic impact of NPFF in rats [ 46 , 47 ] and in chicken [ 48 ] but also NPAF [ 49 ] and NPSF [ 50 ] showed reduced food consumption in chicken. This effect on food consumption is generally described for almost all RFamide peptides in several vertebrates [ 51 , 52 ]. Furthermore, NPFF is known to be involved in water balance [ 47 , 53 , 54 ] and with NPSF and NPAF affects adipogenesis [ 55 ]. Moreover recent investigations in rodents describe NPFF to be involved in locomotion and reward [ 56 - 59 ]. Central injections of NPFF in mice evoked hypothermic effects [ 60 - 62 ] and could be antagonized by BIBP3226 as well as by RF9 [ 63 , 64 ]. No intrinsic affinity at any of the opioid receptors subtypes was observed for NPFF [ 4 , 65 , 66 ], nevertheless NPFF has been recognized as an effective modulator of opioid functions and pain. Additionally, NPFF was found to take part in the development of opiate tolerance and dependence [ 67 , 68 ], albeit the observed anti-/pro-opioid effects of NPFF are dependent on the site of administration [ 38 , 69 ]. PrRP has been initially identified as a prolactin-releasing factor [ 8 ] but follow-up studies rather suggest that the major effects of this neuropeptide are different. In rodents, PrRP is involved in the control of body weight homeostasis inducing anorexia but it has, in contrast to NPFF, no impact on locomotion, exploratory time, grooming and resting time [ 70 , 71 ]. Further studies revealed evidence for a role of PrRP as a regulator for stress response [ 72 - 74 ], nociception [ 74 ], and the cardiovascular system, i.e. arterial blood pressure [ 75 ]. PrRP effects were additionally described to be mediated by the NPFF receptor system in vivo and in vitro [ 76 ] but also via the corticotropin releasing hormone system [ 77 ], which is involved in the mediation of cardiovascular [ 78 ] or anorectic effects [ 79 ]. Due to its implication as a suppressing factor for metastasis, the processed 54 amino acid peptide kisspeptin was originally named metastin [ 10 - 12 , 15 ]. The role of the kisspeptin/metastin encoding Kiss1 gene in cancer metastasis has been recently reviewed [ 80 ]. The common perception of metastin/kisspeptins as pivotal regulator for reproduction [ 81 - 84 ] was the inception for a plethora of studies in this field. Either central or peripheral administration of the peptides kisspeptin-54 and/or kisspeptin-10 were identified to stimulate the gonadotropin-releasing hormone (GnRH), luteinizing hormone (LH), follicle-stimulating hormone (FSH), and testosterone secretion [ 81 - 84 ]. Furthermore, the importance of the kisspeptin receptor GPR54 was elucidated as some cases of hypogonadotopic hypogonadism were associated with mutations of the GPR54 gene [ 85 - 87 ]. Mice, lacking the gene of the receptor fail to undergo sexual maturation [ 86 ]. In contrast to the other RFamide peptides, kisspeptin/GPR54 is not directly involved in body weight control and food consumption [ 86 , 88 ]. It was only described to reduce food consumption by increasing the meal intervals in mice [ 89 ] or the reduced hypothalamic expression of kisspeptin itself was observed for leptindeficiencyanddiet-induced obesity in mice [ 90 ]. The multiple physiological functions of the kisspeptin/kisspeptin receptor signaling as a nodal point in the neuroendorinical regulation of puberty and reproduction [ 91 ] have been recently reviewed and are not focus of the present review [ 92 - 95 ]. The name of the gonadotrophin-inhibitory hormone already describes the initially discovered function for the RFamide related peptide, which was discovered in 2001 in quail brains [ 96 ]. Further exploration identified the mammalian orthologs of the avian GnIH, called RFamide related peptide-1, -2 or -3 (RFRP-1, -2, and -3) and their functional roles. RFRP-3 application in rodents confirmed the inhibitory effects on GnRH neurons [ 97 ] and further investigations suggest an inhibitory modulation of GnRH stimulated LH secretion at the pituitary [ 98 , 99 ]. But RFRP-3 has also a strong influence on feeding and sexual behavior in rats [ 100 ]. In ovine, RFRP-3 was shown to project to cells being involved in regulation of energy balance and reproduction [ 101 ]. For clarification, it has to be emphasized that the human RFRP-3 shares the same sequence with the ovine RFRP-3 and more importantly, it is also designated as NPVF in humans. Previous investigations reported NPVF as NPFF 1 R selective agonist, which attenuate morphine-induced antinociception more effective than NPFF. Furthermore, NPVF is suggested to be important for the endogenous anti-opioid mechanism [ 25 , 102 , 103 ]. The hyperthermic effects of NPVF resemble those of NPFF [ 62 , 63 ]. In chicken, NVPF regulates appetite in a short term manner and its effects are related with hypothalamic and behavior changes [ 104 ]. The NPVF-induced satiety was shown to be mediated through μ and κ but not δ subtypes of opioid receptors in chicken [ 105 ]. The latest member of the RFamide peptides, 26RFa or 43RFa (QRFP) demonstrates a variety of different physiological processes to be involved in, such as regulation of the cardiovascular system by elevating the arterial blood pressure and the heart rate in rodents [ 106 , 107 ]. Administration of either 26RFa or QRFP plays an important role in energy homeostasis by regulating the appetite and energy expenditure, finally resulting in obesity with hyperphagia [ 14 , 107 - 110 ]. Additionally, both ligands possess the ability to activate the gonadotropic axis [ 111 , 112 ] and can increase locomotor activity [ 107 , 109 ]. Recent data present 26RFa able to produce an analgesic effect [ 113 ] and to modulate the nociceptive transmission, finally evoking anti-allodynic effects [ 114 ]. Furthermore, 26RFa is suggested to promote bone formation [ 115 ]. This is suggested by the identification of the 26RFa gene as a potential osteoporosis gene [ 116 ]. A common approach to identify the most important segment of a ligand is to determine the minimal sequence, which is required to activate the receptor. To note, the first described structure-activity study of an RFamide peptide on molluscan muscle revealed that FMRF-OH retained 1/100th of the activity of FMRFamide, while MRFamide had 1/1000th [ 1 ]. Thus, the amidation as well as a certain peptide length had to be considered for future SAR studies at vertebrates. Indeed, the amidated C-terminal fragment of the ligand family is essential for all receptors within this family. Structure-affinity/activity relationship studies can be used to characterize the interaction between ligand and receptor. Essential segments of the peptide and essential segments of the receptor for the interaction can be identified and distinguished from non-essential residues. Furthermore, the investigation of the three-dimensional structure of biologically active peptides is important for the clarification of structure-activity relationships and thus for the development of potent agonists and antagonists [ 117 ]. Many modifications of the NPFF sequence on the N-terminus and the C-terminus have been performed [ 118 , 119 ]. The affinities of NPFF-related peptides inhibiting [ 125 I]1DME or [ 125 I][Tyr 1 ] NPFF specific binding in the dorsal horn of rat spinal cord are summarized in Table 3 . First investigations clearly showed that neither an acetylated N-terminus nor deletion of the two N-terminal residues (not shown) significantly modified peptide affinity [ 118 , 119 ]. In contrast, NPFF(4-8) and NPFF(5-8) showed a 100-fold reduced affinity compared to NPFF. Further on, the carboxy-terminal tripeptide itself (Gln-Arg-Phe-NH 2 ) had only a weak affinity (K i of 300 ± 45 nM) [ 118 , 119 ]. The introduction of the D -enantiomer at position 5 reduced the affinity 100-fold. Accordingly, the most important part is the RFamide motif. Substitutions at position 7 or 8 are not well tolerated and reduced the affinity dramatically. Only substitution to D -Arg in position 7 or Tyr in position 8 showed a moderate loss of affinity about 100-fold compared to NPFF [ 118 , 119 ]. Double substitution of Arg 7 and Phe 8 by the corresponding enantiomer ( D -Arg 7 and D -Phe 8 ) significantly reduced the affinity. The amidated C-terminus is as important as the last two residues. The free acid containing peptide does not show any significant potency compared to the amidated form [ 118 , 119 ]. The classic strategy of amino acid substitution and deletion could clarify the structural requirements for binding to the NPFF receptors. Since then, signal transduction assays were performed in combination with binding experiments on recombinantly expressed NPFF receptor subtypes. The data in Table 4 provide information about the receptor subtype specificity, with regards to the pro-NPFF A and pro-NPFF B derived peptides. These two precursors generate different RFamide peptides. It has been suggested that peptides derived from pro-NPFF A display high affinity/activity at the NPFF 2 R and, conversely, peptides from pro-NPFF B precursor show slight preference for the NPFF 1 R [ 102 ]. Data are similar between the different functional assays and confirm that RFRP-3 (NPVF) and RFRP-1 (NPSF) are more active at the NPFF 1 R than NPFF and NPAF. In accordance with other data, the opposite behavior was found at the NPFF 2 R for NPFF, NPAF and RFRP-3, RFRP-1. Furthermore, shortening the NPFF sequence clearly underlines that these modifications are better tolerated at the NPFF 2 R. And vice versa modifications in the RFRP-3 and RFRP-1 sequence show greater effects at NPFF 1 R [ 28 , 120 , 121 ]. In contrast, by ending up in the tetrapeptide PQRF-NH 2 the same behavior is displayed for both receptors [ 28 , 120 , 121 ]. This is not surprising because of the equal carboxy-terminal four amino acids. As previously reported [ 119 , 120 ], substitutions of the C-terminal Arg and Phe residues result in a significant loss in receptor response. By the use of a systematic approach, we characterized the role of the C-terminal dipeptide with respect to agonistic properties. The synthesized [Xaa 7 ]NPFF and [Xaa 8 ]NPFF analogs were used to investigate signal transduction properties in COS-7 cells, transiently expressing human NPFF receptor subtypes [ 28 ]. Data are summarized in Table 5 . Substitution of Arg with Ala resulted in a peptide that maintained weak activity, but only at high concentrations (> 1 μM) for NPFF 1/2 R. Findeisen et al. [ 28 ] were able to further explore the crucial role of Arg 7 by modifying this position and show that all alterations resulted in identical behavior at the NPFF 1 R and NPFF 2 R: the charge-conserving Lys residue displays the same behavior at both receptors, inducing a ∼100-fold decrease in potency for IP accumulation at NPFFRs. Furthermore, to assess the importance of the cationic guanidinium group within the Arg 7 side chain of the NPFF sequence, terminal NH-groups were monomethylated (MMA) or asymmetric dimethylated (ADMA), resulting in analogs that show an equal activity profile at both NPFFRs as well. No significant change in efficacy was observed for either receptor subtype relative to stimulation with parent NPFF. Interestingly, by decreasing the Arg 7 side chain by either one or two CH 2 groups [replacement by α-amino-4-guanidino-butyric acid (Agb) or α-amino-3-guanidinopropionic acid (Agp)], the resulting analogs exhibit different activity profiles for both receptors. Low potency was observed for both receptors, but the efficacy was different: the [Agb 7 ]NPFF analog was only able to activate the NPFF 1 R up to 31 ± 2% of the wild-type efficacy in contrast to the maximum signal of 96 ± 13% observed for NPFF 2 R. Other hydrophilic amino acids, such as Asp and Cit resulted in dramatically decreased receptor activation. Findeisen et al. postulated that the length of the side chain is important, and that various molecular dynamics are responsible for formation of the agonist-receptor complex in the NPFF 1 R as compared with the NPFF 2 R [ 28 ]. The recently discovered selective NPFF 2 R agonist with a guanidinium group positioned very close to the core structure supports this hypothesis (compounds 1, 3 and 9; Table 10 ) [ 123 ]. These results are in agreement with the requirement of a long, charged side chain with hydrogen bonding potential at this position, suggesting that the specific role of the Arg is critical for best activation of both NPFF receptors. To study the critical role of the C-terminal phenylalanine, a series of [Phe 8 ]NPFF analogs with aromatic and aliphatic amino acids, as well as modifications of the Phe side chain, was investigated by comparing the activity profiles ( Table 5 ). The substitution with alanine led to nearly complete loss of receptor activity for both subtypes, which is in agreement with previous binding studies using the dorsal horn of the rat spinal cord [ 119 ]. Interestingly, the activity was essentially equipotent at both NPFF receptor subtypes after replacement with a cyclohexyl group at position 8 of NPFF and resulted in an approximate five-fold decreased potency. The presence of a hydroxy group in the Tyr side chain resulted in an agonistic peptide with ∼24-fold loss in potency for both NPFFRs. In contrast, substitution of Phe 8 with Trp or His did not generate potent peptides for NPFF 1 R. Both peptides exhibit agonistic activity, but their efficacy was drastically reduced. However, the C-terminal Phe 8 was able to be replaced by Trp or His with only a minor loss in potency at the NPFF 2 R relative to the NPFF 1 R. Still some agonism was observed by the use of methylated Phe analogs. Introduction of the corresponding enantiomer caused a significant loss in potency for both receptor subtypes. Elongation of the side chain by replacement with homophenylalanine (Hph) resulted in a peptide that displayed a significant loss in activity at NPFF 1 R (147-fold) and at NPFF 2 R (444-fold). Remarkably, shortening the Phe side chain to a phenylglycine (Phg) resulted in equipotency for both NPFFRs, although the peptide has decreased efficacy at the NPFF 1 R. With respect to the overall peptide structure, we observed that configuration of the Phe 8 side chain was of similar importance for both subtypes. Additionally, both NPFF receptors were able to be activated by analogs in which the aromatic group is larger than Phe itself (pMePhe). However, only NPFF 2 R retained the full maximum signal when the aromatic ring linker is CH 2 -CH 2 (Hph) or absent (Phg), while repositioning of the phenyl group in relation to the peptide backbone was critical for full NPFF 1 R activation. The results propose that the agonist-receptor complexes for NPFF 1 R are more susceptible to structural modifications ( Table 5 ). Chartrel et al. have isolated a 26 amino acid RFamide peptide from frog brain (hence 26RFa) [ 14 ]. Later, an N-terminally extended form of 26RFa, termed 43RFa or QRFP, was identified and purified from rat brain extract [ 107 ]. Both peptides, 26RFa and 43RFa have been isolated from the human spinal cord and hypothalamus [ 7 ]. The amino acid sequence of the carboxy-terminal heptapeptide of 26RFa has been fully conserved from fish to mammals [ 124 ] and is probably processed in vivo by prohormone convertases. In previous studies it has been demonstrated, that in mammals the amidation of the C-terminus is essential for biological activity of all RFamide-related peptides [ 119 , 125 ]. Previous data showed that the desamidated 26RFa has a 400-fold loss in affinity for GPR103 compared to the amidated analog [ 18 ]. By measuring the potency to induce [Ca 2+ ] i mobilization in Gα 16 -hGPR103-transfected CHO cells, structure-activity studies of a series of analogs revealed that the 26RFa is equally potent as the N-terminally elongated 43RFa, whereas the conserved heptapeptide 26RFa(20-26) was substantially less potent (75-fold), but as efficacious as h26RFa to increase [Ca 2+ ] i [ 126 ]. Further on, by using an Ala-scan the authors could show that the last three carboxy-terminal residues Phe-Arg-Phe are involved in the activation of human GPR103. The conserved heptapeptide can be optimized at position 23, by replacing a serine with a norvaline residue to gain a slightly more potent analog compared to the 26RFa(20-26) ( Table 6 ). They investigated the contribution of the C-terminal carboxamide to the biological activity by using a series of mono- and disubstituted amide 26RFa(20-26) analogs. With these experiments, they could demonstrate that the C-terminal primary amide is involved in receptor binding via a hydrogen bond [ 126 ]. The gradual decline in the biological activity observed with the N-terminally truncated analogs has to be correlated with the affinity of these fragments for GPR103 [ 18 ]. Up to now this is the only activity study, which is published so far. Regarding to the peptide structure, Thuau et al. described the NMR conformation of 26RFa in water and methanol as well as CD spectroscopic studies. Experiments in methanol suggest that the Pro 4 -Arg 17 region of 26RFa adopts a well-defined α-helical structure [ 127 ]. Further on, the helix is flanked by two unstructured regions, a three amino acids short N-terminally one (Thr 1 -Gly 3 ) and a longer segment consisting of the C-terminally conserved moiety (Lys 18 -Phe 26 ). Both, CD- and NMR-studies, reveal a helix with amphiphatic character consisting of a hydrophobic (residues Leu 5 , Leu 8 , Ala 9 and Leu 12 ) and a hydrophilic face (residues Asn 7 , Glu 10 , Glu 11 , Asn 13 and Arg 17 ) [ 127 ]. SAR for the PrRPR and its equipotent agonistic ligands PrRP20 and PrRP31 are quite clear. Early findings revealed that the analog with the C-terminal acid of PRP31 is inactive [ 8 , 125 ]. [ 125 I]-PrRP20 is described as a high affinity ligand, whereas other peptides like NPY, NPFF, RFRP-1 or RFRP-3 show affinities > 30 μM [ 9 , 128 ]. In a down-sizing attempt, Roland et al. identified the heptapeptide PrRP25-31 to be the minimal active agonist fragment and confirmed the importance of the C-terminal amidation for the biological activity of PrRP. Using an Ala-scan they figured out that the three Arg residues at positions 23, 26 and 30 are of importance [ 129 ]. Danho et al. found that Ac-PrRP-(26-31)-hexapeptide was the smallest agonist sequence, and concluded that the critical amino acids were Arg 26 , Pro 27 , Val 28 , Arg 30 and Phe 31 [ 130 ]. The findings of an L-shaped peptide structure with a sloppy N-terminal region and a hydrophobic cluster consisting of Pro 27 ; Val 28 and Phe 31 were confirmed by recent CD- and NMR-studies revealing an amphiphilic helix of the carboxy-terminal region [ 130 , 131 ]. The first comprehensive study on structure-activity relationships at a plethora of PrRP(19-31) analogs was performed by Boyle et al. [ 125 ] ( Table 7 ). The data confirmed the role of the functionally important residues that are located within the carboxy-terminal heptapeptide segment Ile 25 -Arg 26 -Pro 27 -Val 28 -Gly 29 -Arg 30 -Phe 31 -NH 2 , especially of residues Arg 26 , Arg 30 , Phe 31 as well as the C-terminal amide function. Repositioning of the aromatic ring at Phe 31 was not tolerated and substitution of aliphatic or polar residues resulted in little functional activity. In contrast, substitutions of the aromatic ring in correct distance to the backbone were well tolerated. No functional activity was observed for any modification made at position Arg 30 , emphasizing its crucial role, whereas Arg 26 is also relevant but has reduced impact [ 125 ]. Gly 29 is very important for good functional activity and even slight changes are poorly tolerated. Pro 27 is probably required for its turn-promoting property as modifications resulted in less affinity. Val 28 could be replaced by Phg with full retention of functional and binding activity, indicating that it is supplying a reasonably sized hydrophobic side chain in the L-configuration and exhibiting branching close to the peptide backbone. Ile 25 also accepts substitution by Phg but it seems to be a less important position as Ala in position 25 also retained considerable functional activity. The data obtained from a minor number of analogs in positions 21 to 23 confirm that these residues are less important. Backbone methylation from positions 26 to 31 resulted in reduced or none functional activity [ 125 ]. Recently, the investigations at position Phe 31 have been intensified and confirmed the important function of Phe 31 PrRP analogs in vivo and in vitro , with respect to binding, activity, and food intake in fastened mice after central administration [ 132 ]. In detail, analogs with deleted Phe or containing Phe derivatives with bulky side chain or halogenated aromatic ring were tested and revealed high binding potency and cell signaling in RC-4B/C cells [ 132 ]. Since the discovery of the kisspeptin receptor (GPR54) and its full-length ligand Kp-54, numerous peptide and non-peptide ligands have been investigated in structure-activity studies. The smallest highly potent agonist represents the N-terminally truncated Kp-10 and was the lead structure for further structure-activity studies ( Figure 1 ) [ 15 , 133 ]. Later studies revealed that an even minor alteration of the amidated C-terminus is not tolerated at all [ 134 ]. Initial D -amino acid scanning experiments of Kp-10 revealed that the five C-terminal residues are stereochemically of high importance for proper kisspeptin receptor activation [ 135 ]. This is in a line with NMR solution experiments of a Kp-13 in a membrane-like environment (SDS micelles), suggesting a relatively stable, helical conformation from residues 7 to 13 [ 136 ]. Similar results were obtained by molecular modeling, showing that Kp-10 exhibits a helical structure of the C-terminal 7 residues in an α- and 3 10 -characteristic [ 133 ]. In Ala-scanning experiments of Kp-10, the residues Phe 6 , Arg 9 , and Phe 10 resulted in a high loss of agonistic activity [ 133 , 135 , 136 ], which structurally fits well with the structural data, as these three residues lie on one face of the helix and define a pharmacophore site for kisspeptin [ 136 ]. Structure-activity studies at the C-terminal Phe resulted in an improved activity by substitution with Trp ( Table 8 ) [ 136 , 137 ]. Based on this knowledge, Niida et al. developed the significantly down-sized analogs FM052a and FM053a with Kp-10 like agonistic activity ( Table 8 ) [ 135 ]. Extensive optimization of this pentapeptide-based C-terminal kisspeptin analogs resulted in the potent kisspeptin receptor (GPR54) agonist: H-Amb-Nal(2)-Gly-Leu-Arg-Trp-NH 2 (compound 34; Table 8 ) [ 134 ]. Further SAR and QSAR studies, performed on the N-terminal acyl groups of the pentapeptide agonists demonstrated that an aromatic acyl group and the inductive electronegative as well as small substituents at 4-position of the aromatic ring contributed to the agonistic activity. This resulted in the identification of the kisspeptin receptor (GPR54) activating agonist FTM080, which is equipotent with Kp-10 ( Figure 1 ) [ 140 ]. The analog FTM145 was developed to prevent degradation by proteinases under physiological conditions, in which the ( E )-alkene dipeptide isostere at the Gly-Leu site was inserted. FTM145 revealed a higher stability in murine serum and resistance to matrix metalloproteinase mediated cleavage as its half-life was increased up to 38 h, whereas Kp-10 was completely digested after 1 hour [ 139 , 141 ]. Kp-10 exhibited highly potent binding affinity and receptor activation at both NPFF receptors, whereas FTM145 and FTM080 showed reduced bioactivity toward NPFFRs, yet retaining equipotent bioactivity as Kp-10 toward kisspeptin receptor (GPR54). Highly selective agonists are as important as specific receptor antagonists for functional investigations. It should be noted, that Curtis et al. also tested several Kp-10 analogs and describe [ D -Y] 1 Kp-10 to bind to kisspeptin receptor (GPR54) with lower affinity while exhibiting similar bioactivity in vitro . However, peripheral administration of [ D -Y] 1 Kp-10 increased plasma LH and testosterone in vivo more potently than Kp-10 itself in mice. [ D -Y] 1 Kp-10 was suggested to be more stable for proteolytic degradation compared to endogenous Kp-10. This hypothesis was supported by significantly increased total testosterone levels, measured 60 min after injection of 0.15 nmol [ D -Y] 1 Kp-10 whereas the same dose of Kp-10 had no significant effect.

The RFamide peptides represent a family with a strong therapeutic potential as they are involved in numerous regulatory mechanisms related to energy homeostasis, reproduction, pain and behavioral processes like food intake, locomotion and stress response. Their roles in the regulation of hypothalamic functions are summarized in Figure 2 and reflect the multiple mediated effects but more importantly their different impact on these functions. Due to the structural homology of the RFamide peptides they are susceptible to address other GPCRs within this family and evoke a so called crosstalk. This fact is a hurdle in investigations of the distinct roles for single RFamide peptides, e.g. the Kp-10 was described to target the NPFF receptors with low nanomolar affinities [ 138 ] and PrRP effects were found to be mediated through the NPFF receptors as well [ 76 ]. But also the NPFF receptors themselves share high affinities to their endogenous ligands, NPFF and RFRP-3 (NPVF) [ 28 ] and the NPFF 2 R has even been described to bind the Y 1 receptor antagonist BIBP3226. The capability to perform crosstalk is reasonably high for the family of RFamide peptides and thus, selective agonists as well as antagonists are needed to explore the distinct and yet not fully understood mechanism of the evoked diverse pharmacological effects. The development of small ligands might finally lead to small, low molecular weight and lipophilic compounds/ligands as drugs/tools to use them in therapy. As all RFamide peptides take part in fundamental neuroendocrine, behavioral and sensory functions the exploration of selective agonists and antagonists is a fundamental step to gain knowledge of distinct interactions. Antagonists, especially highly selective ones are a tool of choice for the investigation of distinct physiological effects mediated by single receptors and their ligands in vivo. First steps were made in the development of kisspeptin antagonists, which might provide a valuable instrument to elucidate the physiological and pathophysiological role of kisspeptin in the regulation of reproduction and could offer a unique therapeutic agent for treating hormone-dependent disorders of reproduction, including precocious puberty, endometriosis, and metastatic prostate cancer. The most advanced group of antagonistic compounds was developed for the NPFF receptor ligand system, which represents a promising target for drugs with its impact on nociception, food intake and energy expenditure. They may help to optimize the impact of opiates or improve feeding behavior. Up to now, no specific antagonists have been described for the QRFP receptor (GPR103) and the PrRP receptor (GPR10). But it has to be mentioned, that NPY dosed in the micromolar range showed some antagonistic effects for PrRP [ 145 ]. Several putative antagonists have been reported to target NPFF receptors, e.g. compounds derived from endogenous NPFF ligands. Development of antagonists for the NPFFR/GnIH system included N-terminally truncated peptides, like desaminotyrosyl-FLFQRFamide, dansyl-PQRamide and PFR(Tic)amide [ 42 , 146 - 149 ]. Unfortunately, most of them act as partial agonists with a low affinity and thus limit their use as pharmacological tools. In 2002, Mollereau et al. showed that BIBP3226, a prototypical NPY 1 R non-peptidic antagonist [ 150 ], displays high affinities for NPFF receptors [ 120 ]. Since then, several studies demonstrated the antagonistic properties of BIBP3226 at NPFF receptors in vitro and in vivo ( Table 9 ). Nevertheless, its application is limited due to its simultaneous blockade NPY 1 receptors [ 64 , 120 , 151 - 153 ]. It has been suggested that BIBP3226 and related compounds may mediate some of their in vivo effects through NPFF receptors rather than through NPY 1 receptors [ 154 , 155 ]. In 2002, similar compounds to BIBP3226 were described in a patent [ 156 ]. Recently, a derivative of the dipeptide RFamide (RF9) has been suggested to be a potent and selective antagonist [ 157 ]. Selective binding of RF9 to recombinantly expressed receptors in CHO and COS-1 cells was shown (K i -values of 58 nM and 75 nM for NPFF 1 R and NPFF 2 R, respectively) ( Table 9 ). RF9 antagonizes the NPFF-/NPVF-induced agonism in functional assays in vitro and i.c.v. and eliminates NPFF-induced pressure and tachycardiac responses. When RF9 was coinjected with NPFF, it was able to act as an antagonist to block delayed and long-lasting heroin-induced hyperalgesia [ 157 ]. Further on, recent data demonstrate that RF9 prevents NPFF-induced drops of the body temperature and morphine analgesia in mice [ 158 ]. Taken together it is now clear, that NPFF receptors mediate the hypothermia and anti-morphine action of NPFF. Additionally, data revealed a disinhibitory role of NPFF and NPVF in the hypothalamic PVN. The reduction of evoked bicuculline-sensitive inhibitory postsynaptic currents was described, which was blocked by RF9 [ 159 ]. However, in an IP accumulation assay, we observed agonistic activity with full efficacy at both NPFF receptors of transiently transfected COS7 cells after stimulation with RF9 (unpublished data). As former studies of RF9 were tested in cAMP-assays or investigated its activity via stimulation of [ 35 S]GTPγS binding to hNPFF 2 R membranes, it can be speculated that different signaling pathways of the NPFF receptors are responsible for the observed different effects. Recent reports suggested that the different activation of NPFF receptors could generate the diversity of pharmacological effects in vivo [ 62 , 160 ]. It should be noted that a couple of patents have described small nonpeptidergic ligands for NPFF 1 R and NPFF 2 R. In 2003, a patent reported on small templates of quinolino and quinazolino guanidines where different substitutions on the ring system were carried out [ 161 ]. In 2004, derivatives of a thiazole guanidine template with different substitutions have been described as NPFF 1 R antagonists, with IC 50 -values in low nM range [ 162 ]. Unfortunately, functional data for hNPFF 1 R or hNPFF 2 R as well as binding studies on hNPFF 2 R are missing. A later patent refers to the guanine derivatives linked to an aromatic heterocycle to be used as NPFF receptor antagonists [ 163 ]. The most recent patent is based on reference [ 162 ], describing further potent and specific, low-molecular antagonists of neuropeptide FF1 receptors with non-peptide or non-peptoid structures [ 164 ]. Roseweir et al. created some potent peptide antagonists for the kisspeptin receptor (GPR54), based on the substitution of Leu 8 in Kp-10 with D -Trp in combination with Ser 5 substitution by Gly ( Table 9 ) [ 165 ]. The selected antagonist peptide 234 (p234) reduced pulsatile GnRH secretion in female pubertal monkeys and was able to inhibit the firing of GnRH neurons in the brain of the mouse. The inhibitory effect of LH release in rats and mice and the blocking of LH rise in postcastrated sheep, rats, and mice indicate that kisspeptin neurons mediate the negative feedback effect of sex steroids on gonadotropin secretion in mammals [ 165 ]. Pineda et al. extended and refined the in vitro and in vivo testing of the first kisspeptin antagonist p234 by its continuous infusion, and additionally tagged the leading antagonist compound p234 with an N-terminal penetration sequence ( Table 9 ) [ 166 ]. This is predicted to have a higher permeability through the blood-brain barrier and paves the way for new strategy development for systemic antagonism of (or at least part of) the biological actions of endogenous kisspeptins. More recently, small molecule kisspeptin receptor (GPR54) antagonists with a 2-acylamino-4,6-diphenylpyridine scaffold have been reported ( Table 9 ) [ 167 , 168 ]. Kobayashi et al. used a combinatorial chemistry technology to identify a 2-furoyl group to be the most suitable 2-acyl group of all tested 2-acylamino-4,6-diphenylpyridine derivatives, and finally these structure-activity relationship studies led to compound 9l with an IC 50 value of 3.7 nM in a kisspeptin receptor (GPR54) binding assay. Moreover, these compounds showed apparent antagonistic activity in a cellular functional assay [ 167 ]. The optimized compound 15a exhibited high affinity to human and rat kisspeptin receptor (GPR54), apparent antagonistic activity, and high brain exposure. In addition, intravenous administration of 15a to castrated male rat suppressed the plasma LH level, which indicates the possibility of a small molecule kisspeptin receptor (GPR54) antagonist as a novel drug for sex-hormone dependent diseases [ 168 ]. Gaubert et al. have reported the discovery and characterization of the first nonpeptidic selective NPFF 2 R agonists. They demonstrated that systemic administration of the selective NPFF 2 R agonists (compound 1 and 3; Table 10 ) were active in various pain models in vivo , whereas administration of a nonselective NPFF 1 R and NPFF 2 R agonist (compound 9; Table 10 ) increases sensitivity to noxious and non-noxious stimuli [ 123 ]. By performing two functional assays (R-SAT and cAMP assays) in addition to binding assays using 125 I-NPFF, Lameh et al. investigated the activity profiles of several novel nonpeptidic, small molecule NPFF agonists for recombinant human NPFF 1 and NPFF 2 receptors ( Table 10 ). In this study, they could identify compounds that, according to the performed in vitro functional assays, are selective agonists for NPFF 2 R ( i.e. , AC-263093) and nonselective agonists (AC-262616) for NPFF receptors ( Table 10 ). Additionally they found AC-262620 and AC-262970 to be selective NPFF 1 R antagonists, which bind NPFF 1 R with low nanomolar affinity but behaved as full agonists at NPFF 2 R [ 169 ]. Using these compounds they were able to clarify the pharmacology of NPFF receptors after systemic administration. Further on, the in vivo results provide evidence for the divergent roles of NPFF receptor subtypes in the modulation of nociception. They could demonstrate that NPFF 1 R activation is pronociceptive, whereas NPFF 2 R activation is antinociceptive [ 169 ].