A Novel and Selective -Melanocyte-Stimulating Hormone-Derived Peptide
http://www.100md.com
《内分泌学杂志》
Divisions of Endocrine Research (H.M.H., J.H., X.Z., D.P.S., M.L.H.), Bio-therapeutics Discovery Research (D.L.S., D.D.Y., J.P.M., L.Z., L.Z.Y.), Discovery Chemistry and Research Technology (H.M.)
Lead Optimization Biology (S.H.), Eli Lilly & Co., Indianapolis, Indiana 46285
Abstract
MSH has generally been accepted as the endogenous ligand for melanocortin 4 receptor (MC4R), which plays a major role in energy homeostasis. Targeting MC4R to develop antiobesity agents, many investigators have performed a structure-activity relationship (SAR) studies based on MSH structure. In this report, we performed a SAR study using human MSH (5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ) (DEGPYRMEHFRWGSPPKD, peptide 1) as a lead sequence to develop potent and selective agonists for MC4R and MC3R. The SAR study was begun with a truncation of N terminus of MSH (5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ) together with acetylation of the N terminus and amidation of the C terminus of the peptide. Introduction of a cyclic disulfide constrain and replacement of L-Phe with D-Phe afforded a super potent agonist (peptide 5). Furthermore truncation at the C terminus generated a small and potent MC4R and MC3R agonist (Ac-YRcyclo[CEHdFRWC]amide, peptide 6), which exhibited no MC5R and greatly reduced MC1R activity. Molecular modeling of Ac-YRcyclo[CEHdFRWC]amide (peptide 6) revealed that Arg2 in the peptide formed a salt bridge with Glu4. Subcutaneous or intracerebroventricular administration of peptide 6 in rats showed potent in vivo efficacy as evidenced by its effects in reducing energy balance, increasing fat use, and decreasing weight gain in both acute and chronic rat metabolic studies. Furthermore, the antiobesity effect by peptide 6 was manifested only in wild-type but not MC4R-deficient mice, indicating that antiobesity effects of the peptide were attributed largely through MC4R but not MC3R agonist activity of the peptide.
Introduction
THE PROOPIOMELANOCORTIN (POMC) gene expression in tissues produces a 36-kDa preprohormone, which is cleaved by a signal peptidase to a 32-kDa prohormone, POMC (1). POMC protein (32 kDa) is further processed in a tissue-specific manner by various processing enzymes to generate multiple and diverse peptide hormones including MSH, MSH, MSH, ACTH, -lipotropin (LPH), and -endorphin (1). Specifically, proPOMC protein is processed first by prohormone convertase (PC)1 in tissues to generate proACTH (22 kDa) and LPH (10 kDa). ProACTH is further processed to form N-terminal POMC peptide and ACTH (4 kDa). ACTH is then processed further by PC2, carboxypeptidase E, peptidyl -amidating monooxygenase, and N-acetyltransferase to generate mature MSH and des-acetyl-MSH (1, 2). In most mammalian species, LPH is processed by PC2 to generate MSH and -endorphin. However, due to the lack of dibasic residues at the MSH processing site in the rodent LPH sequence, mature MSH cannot be generated in rodents (1, 3). Furthermore, because PC2 is not present in the human pituitary gland, the main POMC-derived peptides in human pituitary are N-terminal POMC peptide, ACTH, and LPH whereas MSH and MSH are produced in hypothalamus, brain, and some peripheral tissues (1).
Diverse lines of evidence, including genetic and pharmacological data obtained in rodents and humans, suggest that the melanocortin 3 receptor (MC3R) and MC4R play important roles in the regulation of energy homeostasis in animals (4). In fact, MC4R mutation is the most prevalent single gene mutation that caused human obesity (5, 6). The human and rodent genetic data strongly support that MC4R is the most validated target for obesity and the POMC-derived peptides are native ligands for MC4R and MC3R. Among the POMC-derived peptides, MSH was long presumed to be the primary ligand for MC4R (7), and MSH was regarded as the primary ligand for MC3R (8). On the other hand, MSH’s role with respect to the MCRs is not fully understood, and it is not considered to be a major or important ligand for any of these receptors.
Although MSH was considered to be the primary peptide ligand for MC4R and the peptide was shown to decrease feeding in rodents when injected intracerebroventricularly (ICV) (9), MSH binds to cloned human MC4R or rat hypothalamic homogenate with only moderate affinity [inhibitory constant (Ki) = 324 and 22.5 nM, respectively] (10). Additionally, Harrold et al. (11) reported that MSH concentration in the hypothalamus was not changed, whereas the concentration of an endogenous MC4R and MC3R antagonist, agouti-related protein, was altered by changes in the nutritional status of the animals. These data suggest that, in addition to MSH, perhaps other peptide or protein ligands may also play a role in regulating energy homeostasis by interacting with MC4R and MC3R in hypothalamus.
Several recent reports suggested that MSH may serve as a particularly important endogenous ligand for MC4R and MC3R in the regulation of energy homeostasis. First, MSH bound cloned human MC3R and MC4R with equal or higher affinity when compared with MSH (10, 11, 12, 13). Second, MSH is produced at relatively high concentrations in the hypothalamus (1, 10, 14), a key region for regulating energy homeostasis. Finally, some POMC mutations associated with human obesity are localized in the coding region or processing sites of the human MSH gene (13, 15). These mutated POMC genes would generate a MSH--endorphin fusion protein and truncated MSH peptides. The MSH--endorphin fusion protein was shown to be a potent antagonist to MC4R. Therefore, the mutant protein can disrupt the function of the native ligand (MSH) with MC4R by a dominant-negative effect (13). Alternatively, some human mutations that produced truncated MSH or no MSH also exhibited obese phenotype (15). These combined data suggest that the MSH peptide could play a physiologically important role in regulating energy homeostasis.
Early melanocortin peptide structure-activity relationships (SARs) studies focused mostly on MSH and its analogs (16, 17, 18). As a result, several potent but nonselective melanocortin peptides, including (Nle4, D-Phe7) (NDP)-MSH, MTII [Ac-Nle-cyclo (DHdFRWK)amide] and SHU9119 [Ac-Nle-cyclo(DHdNalRWK)amide] were developed as important tools to study MCR pharmacology. The underlined and italicized amino acid sequences denote the core peptide sequences for MCR binding. In this report, we describe a focused SAR study using human MSH(5–22) as a lead sequence. Our SAR study, based on the MSH structure, generated several potent and selective MC4R and MC3R peptide agonists. Additionally, we observed that a potent and selective MC4R and MC3R peptide agonist, Ac-YRcyclo[CEHdFRWC]amide (peptide 6 in Table 1) reduced food intake and body weight in diet-induced obese (DIO) rats and normal chow-fed rats. Finally, we demonstrate that the effect of the peptide 6 was mediated through its MC4R activity as shown by its lack of activity in MC4R-deficient mice.
Materials and Methods
Peptide synthesis
All peptides were synthesized by solid-phase methods (19). The solid supports were either preloaded Wang resin for C-terminal acid or MBHA Rink resin for C-terminal peptide amide (20, 21). The primary peptide chain was assembled using an ABI 433A synthesizer (PE Applied Biosystems Inc., Foster City, CA). Side-chain protection of amino acids were compatible to standard Fmoc chemistry, as shown below: Arg(Pbf), Asp(OtBu), Cys(Trt), Glu(OtBu), Gln(Trt), His(Trt), Lys(Boc), Ser(tBu), Thr(tBu), Trp(Boc), Tyr(tBu), or otherwise specified. Other common and uncommon amino acids were used without side-chain protection. Single coupling of each residue was performed using dicyclohexylcarbodiimide/hydroxybenztriazole activation protocol, with a 4-fold excess of each amino acid and 1.5 h coupling time. Acetylation of the -amino group after the chain assembly was normally carried out off-line with 5 eq acetic anhydride, 10 eq N,N-diisopropylethyl amine in dry N-N-dimethylformamide for 1 h at room temperature. The finished peptide was simultaneously deprotected and cleaved from the resin using a cocktail of trifluoroacetic acid/H2O/triisopropyl silane (Tis)/1,2-ethanedithiol (95/2/1/2, vol/vol), or trifluoroacetic acid/H2O/Tis/anisole (92/2/4/2, vol/vol) for 2 h at room temperature. The solvents were then evaporated under vacuum, and the residue was washed three times with cold ether to remove the scavengers. The residue was redissolved in aqueous acetonitrile and purified on reverse phase-HPLC. In case of peptides with disulfide bond, the crude product was dissolved in dimethylsulfoxide and then diluted with 0.2 M ammonium acetate buffer (pH 7) to a final concentration of 20% of dimethylsulfoxide. The solution was agitated at room temperature overnight to form the disulfide bond. Then the solution was diluted with two volumes of water and purified by a preparative RP-HPLC. Depending on the amount of peptide to be purified, different size of the reverse-phased column (Vydac or Zorbax C18, diameter ranges from 0.9 to 5.0 cm) was chosen accordingly. The fractions containing the desired product were pooled and lyophilized. Further characterization of the final product was performed using analytical HPLC and mass and UV spectrometry.
MCR binding assays
We performed radioligand-binding assay by using 125I-NDP-MSH (sp act. 2000 Ci/mmol; Amersham, Piscataway, NJ) as the radioactive ligand. Cell membranes were prepared from the human embryonic kidney 293 cells stably transfected with cloned human MCRs. Cells were grown as adherent monolayers in roller bottle cultures at 37 C and 5% CO2/air atmosphere in a 3:1 mixture of DMEM and Ham’s F12 containing 25 mM l-glucose, 100 U/ml penicillin G, 100 μg/ml streptomycin, 250 ng/ml amphoterin B, 300 μg/ml gentisin, and supplemented with 5% fetal bovine serum. For large-scale production, monolayer cells are adapted to suspension culture (22) and are grown in either spinner or shaker flasks (37 C and 7.5% CO2/air overlay) in a modified DME/F12 medium containing 0.1 mM CaCl2, 2% equine serum, and 100 μg/ml sodium heparin to prevent cell-cell aggregation. Cells are harvested by centrifugation, washed in PBS, and pellets stored frozen at –80 C.
For the preparation of membranes, frozen cell pellets were resuspended in 10 volumes of membrane prep buffer (10 ml buffer per gram of cell paste) consisting of 50 mM Tris (pH 7.5) at 4 C, 250 mM sucrose, 1 mM MgCl2, Complete EDTA-free protease inhibitor tablet (Roche Applied Science, Indianapolis, IN), and 24 μg/ml DNase I (Sigma, St. Louis, MO). The cells were homogenized with a motor-driven Teflon-glass Dounce tissue homogenizer (Wheaton Science Products, Millville, NJ) using 20 strokes, followed by centrifugation at 38,000 x g at 4 C for 40 min. The pellets were resuspended in membrane prep buffer at a concentration of 2.5–7.5 mg/ml, aliquoted, quick frozen in liquid nitrogen, and stored at –80 C.
Standard competitive receptor/ligand binding experiments were performed in 96-well plate formats, and the assay mixtures consisted of serial dilutions of test compounds (10 μM to 100 pM) or those of unlabeled NDP-MSH (100 nM to 1 pM) in binding buffer [25 mM HEPES (H 7.5); 10 mM CaCl2; 0.3% BSA]. The incubation mixture also contained 0.5–5.0 μg membrane proteins, 100 pM 125I-NDP-MSH, and 0.25 mg of wheat germ agglutinin SPA beads. The mixture solutions in 96-well plates were then agitated briefly on a plate shaker and incubated for 10 h at room temperature. The radioactivity bound to the receptor was quantified in a Trilux microplate scintillation counter (PerkinElmer Life Sciences, Norwalk, CT). Nonlinear regression analysis of competitive binding assay data using a four-parameter logistic fit yielded IC50 values, which were converted to affinity constants obtained from competitive binding assays (Ki values). These conversions were performed by using the Cheng-Prusoff equation, Ki = IC50/(1 + D/Kd), where D is the concentration of radioligand and Kd is the equilibrium dissociation constant determined from saturation binding analysis.
In vitro MCR functional assays
Functional activity was determined using a standard cAMP assay with serial dilutions of test compound (10 μM to 0.1 nM) or the control agonist NDP-MSH (100 nM to 1 pM). The human embryonic kidney 293 cells stably transfected with the human MC1R, MC2R, MC3R or MC4R were grown in DMEM containing 10% fetal bovine serum and 1% antibiotic/antimycotic solution. On the day of the assay, the cells were dislodged with enzyme-free cell dissociation solution and resuspended in cell buffer (Hanks’ balanced salt solution without phenol red Hanks’ balanced salt solution-092, 0.1% BSA, 10 mM HEPES) at 1 x 106 cells/ml. Cell suspension (40 μl) was added to positron emission tomography 96-well plates containing 20 μl of diluted compound or control agonist. Plates were incubated at 37 C for 20 min, and the assay was stopped by the addition of 50 μl of quench buffer (50 mM Na acetate, 0.25% Triton X-100).
Per the manufacturer’s (Amersham) instructions, cAMP concentrations were determined by a scintillation proximity assay-based competition assay using 125I-cAMP (Amersham), goat anti-cAMP antibody (MP Biomedicals, Irvine, CA), and polyvinyl toluene antisheep antibody binding beads (Amersham). The assay buffer contained 50 mM sodium acetate and 0.1% BSA. A mixture containing scintillation proximity assay beads (1 mg/ml), antibody (0.65%), and radioligand (61 pM) was prepared in assay buffer, and 100 μl of this solution were added to each well of the 96-well assay plate to yield a final volume of 210 μl. After a 12-h incubation, the plates were counted in a Trilux microplate scintillation counter (PE Life Sciences). The data were converted to picomoles of cAMP using a standard curve obtained from the same assay performed with varying concentrations of unlabeled cAMP. The data were analyzed using a four-parameter logistic nonlinear regression to generate agonist potencies (EC50) and percentages of efficacy relative to the maximum stimulation obtained with NDP-MSH.
Animal and maintenance
All animal experiments were conducted in accordance with principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. In addition, Lilly Research Laboratories Animal Care and Use Committee approved all animal protocols used in these studies. Male Long-Evans rats were weaned and maintained on a diet comprised of 40% fat, 40% carbohydrate, and 20% protein (Teklad 95217; Harlan, Indianapolis, IN) and maintained at Harlan Sprague Dawley (Indianapolis, IN) until 90 d of age when they were used. Rats (500 g) were acclimated (single housed) for at least 1 wk in a temperature- (25 C) and light-controlled (12-h light, 12-h dark; lights on at 2200 h) environment. Water and food were available ad libitum.
ICV cannulation and administration
Male Long-Evans rats (Harlan) approximately 275 g (60 d old) on a normal chow diet (Purina 5008) were implanted with stainless steel cannulae guides (5 mm length, 26 gauge; Plastics One, Roanoke, VA) stereotaxically implanted in the lateral ventricle anteroposteriority: 0.8 mm caudal to bregma; and lateral: 1.5 mm from midline suture. Animals were handled and weighed daily and were placed in calorimeter chambers on the day previous to the test day to minimize stress. On the test day, animals were injected between 1300 and 1400 h with 5 μl of the peptide 6 (0.187 nmol) or saline, and indirect calorimetry was performed for 22–24 h after injection as described below.
Indirect calorimetry
Indirect calorimetry experiments were performed following a published procedure (23). Specifically, 24-h caloric expenditure and respiratory quotient (RQ) were measured by using an open circuit system (Oxymax; Columbus Instruments, Columbus, OH). RQ was measured as the ratio of volume of CO2 produced to the volume of O2 consumed. Caloric expenditure was calculated as the product of calorific value of oxygen and volume of O2 consumed per kilogram of body weight, where calorific value of oxygen = 3.815 + 1.232 (RQ). Total calories expended were calculated to determine daily fuel use. To calculate the proportion of protein, fat and carbohydrate that used during the 24-h period, we used Flatt’s proposal and formula as other derived constants (24). Locomotor activity was measured by counting number of new beam breaks (ambulatory) and same beam breaks (fine movement) during each period of 24 h in the calorimeter. Comparisons between groups of data (mean ± SEM) were analyzed using SigmaStat (SPSS Inc., Chicago, IL) ANOVA using Student-Newman-Keuls (Student-Newman-Keuls) test for multiple comparisons.
MC4R-deficient mice
Generation of MC4R-deficient mice is described as follows: murine genomic DNA clones corresponding to the MC4R locus were isolated from a mouse strain 129/SvJ (Fix II phage library, Stratagene Inc., La Jolla, CA) using a probe generated by PCR amplification of murine MC4R coding sequences. A gene-targeting backbone vector pGT-NN-tk was constructed as described (25). The targeting vector pKO-MC4R contained a 3.2-kb HindIII-NcoI fragment obtained from the 5' end of a MC4R genomic clone and a 1.4-kb HindIII-SacI fragment obtained from the 3' end of a MC4R genomic clone using pGT-NN-tk as a backbone. The gene targeting construct replaced the entire MC4R gene (an internal 1.5-kb NcoI-HindIII genomic fragment) with the neomycin resistance cassette. Electroporation of the linearized pKO-MC4R plasmid DNA into R1 embryonic stem (ES) cells, selection of G418 and ganciclovir-resistant clones, and generation of chimeras were performed as described (26). Male chimeras were bred to female C57BL/6 mice to generate F1 heterozygous mice. Germline transmission of MC4R mutation was confirmed by Southern analysis of tail DNA. The 10th-generation mice of C57BL/6 background heterozygous for the disrupted MC4R allele were bred to generate MC4R-null and wild-type (WT) littermates (Taconic, Hudson, NY). Heterozygous MC4R(+/–) mice were interbred to generate MC4R(–/–) and WT littermates at Taconic Farm. MC4R(–/–) and WT littermates were weaned at 3 wk onto normal chow diet (NIH-31; Ziegler, Gardners, PA) and shipped to Lilly Research Laboratories at 8 wk. Upon arrival the mice were housed individually and permitted to eat (TD 95217; Teklad, Madison, WI) and drink ad libitum. Animals were acclimated for at least 2 wk in a temperature- (25 C) and light-controlled (12-h light, 12-h dark; lights on at 2200 h) environment.
Acute metabolic study
Male Long-Evans rats.
Rats (500 g) were acclimated (single housed) for at least 1 wk in a temperature- (25 C) and light-controlled (12-h light, 12-h dark; lights on at 2200 h) environment. Water and food were available ad libitum. Groups of four rats each were fasted for 4 h and treated once between 1300 and 1400 h by sc injection with vehicle (saline) or peptide 6 (0.3, 1.5, and 3 μmol/kg). The rats were placed in sealed chambers to determine acute changes in energy expenditure (EE) and respiratory exchange ratio by indirect calorimetry (Oxymax; Columbus Instruments International). Rats were permitted to eat a high-fat diet containing 20% protein, 40% fat, and 40% carbohydrate (TD 95217; Teklad) and drink ad libitum throughout the study.
Male MC4R-deficient mice.
Groups of WT (n = 4) and MC4R-null mice (n = 4) were treated once between 1000 and 1100 h by sc injection with vehicle (saline) or peptide 6 (0.3, 3, or 30 μmol/kg). The mice were placed in sealed chambers to determine acute changes in EE and respiratory exchange ratio by indirect calorimetry (Oxymax; Columbus Instruments). Mice were permitted to eat a high-fat diet containing 20% protein, 40% fat, and 40% carbohydrate (TD 95217; Teklad) and drink ad libitum throughout the study.
Chronic metabolic activity study
Rats were randomized to three groups and each group was comprised of five animals. The animals were housed individually throughout the study and permitted to eat (TD 95217; Teklad) and drink ad libitum throughout the study.
Change in total body adipose and lean mass during treatment was measured by nuclear magnetic resonance (NMR) (Echo Medical Systems; Houston, TX) (27) before treatment and on d 15 of treatment. The wide-line NMR signal was calibrated using chicken breast muscle (fat and skin removed) for lean mass and canola oil for fat mass (27). Initial measurement of body composition by NMR was performed 1 d before initiation of treatment and was measured again on d 15. All treatments began on d 1 at 0900 h and continued for 14 d. Groups of five rats each were treated daily by sc injection with vehicle (saline, 1 ml/kg) or peptide 6 (0.3 and 3 μmol/kg). The rats were placed in sealed chambers to determine acute changes in EE and RQ by indirect calorimetry (Oxymax) on d 1, 7, and 14. Animals were then returned to their home cages after each calorimetry session, and daily treatment continued. Body weight and food intake were measured daily at 0900 h.
Serum hormone measurements (insulin, ghrelin, and leptin) and blood chemistry (glucose and corticosterone) for the rats treated with vehicle (saline, 1 ml/kg) or peptide 6 (0.3 and 3 μmol/kg) were performed by Linco Diagnostic Services, Inc. (St. Charles, MO). The rats (five rats per group) were treated in the same manner as described above except the treatment period was shortened to 10 d and the rats were then killed for serum collection.
Statistical analysis
All data are shown as the mean SEM, and statistical comparisons were made using ANOVA followed by Student-Newman-Keuls test when significant (P < 0.05) interaction was found and more than two groups were being compared.
Model structure of Ac-YRcyclo[CEHdFRWC]amide (peptide 6)
Model structure of peptide 6 (Table 1) was built using Quanta98 (Accelrys, San Diego, CA). Energies, including contributions from Van der Waals and electrostatic terms, were minimized in vacuum. Turning on and off explicit hydrogen bonds had no significant impact on the model structure because the cyclic nature of this molecule greatly limits the freedom of side chain placement.
Results
SAR studies
Our previous studies with synthetic melanocortin peptides (12) demonstrated that human MSH(5–22) (peptide 1) and particularly its diaminopeptidase (I and IV) cleaved products are potent but nonselective MC4R agonists. Aiming to develop a more selective agonist for MC4R and MC3R, we performed a further SAR study (Table 1) using human MSH(5–22) DEGPYRMEHFRWGSPPKD (peptide 1) as a lead sequence. Human MSH(5–22) (peptide 1) is a moderately potent agonist with Ki = 23 nM and EC50 = 9.3 nM for MC4R (Table 1). By removing two and four amino acid residues, respectively, from the N terminus of human MSH(5–22), we obtained two nonselective, yet more potent, MC4R peptide agonists, MSH(7–22) (peptide 2, Ki = 4.6 nM) and MSH(9–22) (peptide 3, Ki = 5.7 nM) (Table 1). The binding affinity and functional potency of these two peptides toward MC4R and MC3R were severalfold greater than their predecessor peptide, MSH(5–22).
To increase the stability of the peptide against serum proteases, we further modified the structure of the peptide MSH(9–22) by capping the N terminus with an acetyl group, converting the C terminus acid to an amide, and replacing L-Phe with D-Phe. With these changes, we obtained a super potent but nonselective peptide agonist, peptide 4 (Table 1) for MC1R, MC3R, MC4R and MC5R. Our result with the replacement of L-Phe by D-Phe residue in MSH(5–22) peptide was consistent with the previous literature results with the same modification of MSH-derived peptides (17, 28). This D-Phe replacement not only increases proteolytic resistance of the peptide but also enhances its binding affinity to MC4R by stabilizing the turn structure (17). Aiming to further increase selectivity of the MSH analogs, we altered the peptide structure by incorporating two cysteines in the molecule and generated a cyclic MSH(9–22) analog with a disulfide bond similar to what was previously reported in MSH-derived peptides (17). The resulting cyclic peptide 5 (Table 1) showed a more selective and potent agonist activity for MC4R with no detectable MC5R agonist activity (Ki for MC4R = 0.5 nM and Ki for MC5R > 500 nM). Additionally, removal of the last five residues (SPPKD) from the C terminus of peptide 5 afforded peptide 6 (Table 1), which was a highly potent MC4R agonist (Ki = 0.8 nM; EC50 = 0.6 nM) but showed a much reduced binding affinity toward MC1R (Ki = 17.8 nM). In fact, peptide 6 (MC1R, Ki = 17.8 nM) showed a 30-fold lower MC1R binding affinity than its predecessor peptide, 5 (MC1R, Ki = 0.6 nM).
The cyclic disulfide peptide structure appears to be crucial for its MC4R potency and selectivity. Disrupting the ring structure by replacing cysteines with -aminobutyric acid in peptide 6 resulted in linear peptide 9 that showed significantly compromised MC4R activity and selectivity (Table 1). The D-Phe residue in the center of the sequence appeared to be essential for its affinity toward MC4R. Replacing D-Phe in peptide 6 with L-Phe afforded peptide 10, which exhibited greatly reduced MC4R binding affinity (Ki = 30.5 nM), compared with peptide 6 (MC4R, Ki = 0.8 nM). Furthermore, cyclic peptide 10 also showed much weaker binding activity toward MC1R, MC3R, and MC5R (Table 1).
The SAR data (Table 1) also showed that peptide 6 appeared to reach an optimal MC4R activity or potency. Removal of tyrosine from the N terminus of peptide 6 resulted in peptide 7, which exhibited slightly more potent binding affinity and functional activity for MC4R (Ki = 0.4 nM, EC = 0.4 nM) and MC3R (Ki = 39.4 nM). The exocyclic basic amino acid residue (Arg) present in the MSH(5–22)-derived peptides such as peptides 6 and 7 appeared to be crucial for its potency and activity toward MC4R and MC3R. Additional removal of Arg residue from amino terminus of peptide 7 generated peptide 8 (Table 1), which had weaker binding affinity and functional activity for MC4R (Ki = 2.5 nM, EC50 = 3.3 nM) and MC3R (Ki = 478 nM) than peptides 6 or 7. This less potent MC4R peptide 8 has been previously reported to be an active melanotropin in frog or lizard skin bioassay (29) and a neuroactive substance causing grooming behavior in rats after ICV injections (30).
Acute energy balance changes in DIO rat models administered with peptide 6
Various dosages (0.3, 1.5, and 3 μmol/kg) of peptide 6 together with saline control were administered sc into the male Long-Evans rats maintained on a caloric-dense diet. Subcutaneous administration of peptide 6 (0.3, 1.5, and 3 μmol/kg) in rats reduced food intake to 57, 45, and 25% of the levels of the vehicle control rats treated with saline (Fig. 1A). It was also observed that those rats treated with peptide 6 showed significant increases in fat use [126, 157, and 157% of those of the vehicle controls (Fig. 1B)]. Together we demonstrated that the peptide-treated rats exhibited a negative energy balance in a dose-dependent manner. Treatment with peptide 6 also decreased RQ in DIO rats when compared with vehicle control (Fig. 1C). Additionally, we observed no significant changes in total EE or locomotor activity in the rats treated with the peptide when compared with those of saline-treated rats (Fig. 1, D and E).
Chronic energy balance changes in body weight and lean and fat mass of the DIO rats
Ac-YRcyclo[CEHdFRWC]amide (peptide 6, Table 1) was administered to male Long-Evans rats once daily for 14 d at two different dosages (0.3 and 3.0 μmol/kg). During this period, we measured the changes in body weight and body composition of the rats. Figure 2A shows that peptide 6 at a dose of 3 μmol/kg caused a significant decrease in body weight throughout the 2-wk testing period. Administration of the peptide at 10-fold lower dose (0.3 μmol/kg) also caused significant decrease in body weight at the end of the 2-wk testing period. Furthermore, we observed a gradual rebound in body weight after the administration of the peptide was stopped (Fig. 2A).
Additionally, Fig. 2B shows that rats treated with peptide 6 (3 μmol/kg) exhibited a significant reduction (>20 g) in fat mass with no change in lean mass when compared with saline vehicle controls. We also observed a trend of decreases in serum leptin or triglyceride concentration in the rats treated for 10 d with a high dose of peptide 6 (3 μmol/kg) when compared with those of vehicle controls. However, no changes of serum insulin, glucose, ghrelin, or corticosterone were observed in all treated rats (data not shown).
Acute ICV administration of peptide 6 in rats fed regular chow
To further test the central melanocortin activity of the peptide, we administered the peptide 6 ICV at a very low dose (0.187 nmol/rat) to rats fed on a regular chow. The results (Fig. 3) showed that ICV administration of peptide 6 at a low dose reduced food consumption to 68% (Fig. 3A) and increased fat use to 439% of those observed in vehicle controls (Fig. 3B) over a 24-h period of time. Indirect calorimetry also showed a significant reduction of RQ in the rats treated with peptide 6 when compared with the vehicle control rats (data not shown).
In vivo activity of peptide 6 in WT and MC4R-deficient mice
To further elucidate the mechanism of in vivo activity of peptide 6, we administered this peptide to the MC4R-deficient (MC4R–/–) mice and their sibling WT controls. The results (Fig. 4A) showed that sc administration of peptide 6 at a high dose (30 μmol/kg) to the MC4R-deficient mice showed no effects on food consumption over a 24 h period; however, the treatment of peptide 6 in WT sibling control mice caused a dramatic decrease in food intake (to 19% of WT vehicle control). Furthermore, the results (Fig. 4B) also showed the dose-dependent increases of fat use (158, 203, and 221% of the vehicle control) in WT mice treated with peptide 6. On the other hand, the peptide caused no statistically significant changes in fat use in the MC4R–/– mice (Fig. 4B). These results demonstrate that the major antiobesity effect of peptide 6 was mediated through an MC4R mechanism and not through MC3R activity of the peptide.
Model structure of peptide 6
The model structure of peptide 6 (Fig. 5; first residue as Tyr1) showed that exocyclic Tyr1 residue is well separated from the cyclic core structure that is generally believed to be the main recognizing element in the receptor binding (31). As expected from this model, peptide 7, which was resulted from removal of N-terminal Tyr1 from peptide 6, retains its potency for MC3R and MC4R. However, the second exocyclic residue in peptide 6, Arg2, was shown to interact with units within the core structure, i.e. the side chain of Arg2 forms a strong salt bridge with the side chain of Glu4 (Fig. 4). This ionic interaction may stabilize the main cyclic structure in addition to the stabilization afforded by D-Phe6. It was interesting to note that the side chain of Glu4 was pulled away from the cyclic core structure, presumably due to formation of this salt bridge. If bothTyr1 and Arg2 were removed, the resulting compound, i.e. Ac-cyclo[CEHdFRWC]amide (peptide 8, Table 1), would lose Arg2-Glu4 charge interactions and have flexible side chains. We speculated that the cyclic core structure with its two positive charges (His3 and Arg7) could be the key recognition elements for a negatively charged pocket in MCRs (32). Absence of Arg2 would not only potentially weaken ligand-receptor electrostatic interaction but also make binding entropically unfavorable. Indeed, the Ki values of peptide 8 for MC4R and MC3R were reduced by severalfold (MC4R Ki = 2.5 nM and MC3R Ki = 478 nM, respectively) as compared with those of peptide 6 (MC4R Ki = 0.8 nM, MC3R Ki = 56.4 nM). Thus, Arg2 may have two roles: it neutralizes the Glu4-negative charge so that the positive charges (His3 and Arg7) in the core domain can be recognized by the receptors through charge interaction, and it limits Glu4 side-chain freedom (reduces Glu4 entropy in unbound state) and stabilizes its conformation.
The role of D-Phe provided extra stability to the -turn for the cyclic peptide. Our model structure showed the side chain of His3 might have -stack-like interactions with the side chain of D-Phe6. Replacing D-Phe with L-Phe would be less likely to have such an effect. Furthermore, an electrostatic surface potential plot (not shown) generated from our model shows that Trp8 and D-Phe6 form a hydrophobic extrusion, whereas His3 and Arg7 on the other side have positive charges exposed. Our model supports a recent report (33) that both the hydrophilic surface and the hydrophobic surface are needed for tight binding to the melanocortin receptors.
Discussion
Central MC4R and MC3R are known to play pivotal roles in regulating energy homeostasis in animals. The endogenous ligand for MC4R was long presumed to be MSH, a 13-amino acid peptide hormone produced in hypothalamus, brain, pituitary, and many other peripheral tissues. A primary ligand for MC3R has been hypothesized to be MSH, which has shown selective activity toward MC3R (8). However, MSH such as human MSH(5–22) (DEGPYRMEHFRWGSPPKD, peptide 1, Table 1) isolated from human hypothalamus, was not considered to be an endogenous ligand for MC4R or MC3R, even though it has moderately potent agonist activities for MC3R and MC4R (Table 1).
Bertagna et al. (14) showed that human MSH(5–22) is a true endogenous peptide isolated from human hypothalamus, whereas previously reported 22-amino acid human MSH(1–22) peptide (AEKKDEGPYRMEHFRWGSPPKD) was generated artificially from acid extraction of the peptides from human pituitary. Recently Harrold et al. (10) and we (12) have also proposed that MSH in the hypothalamus is an endogenous MC4R ligand whose activity may play an important role in regulating energy homeostasis. Finally, MSH(1–22) is also a proteolytic fragment corresponding to residues 35–56 of human LPH (34, 35).
That MSH may play an important physiological role in human energy homeostasis has been further supported by human genetic studies (13, 36). Human heterozygous POMC gene mutation (Arg236Gly) causing obese phenotype has been reported at the junction of MSH--endorphin processing site of the gene (13). The patient can produce native MSH but is unable to process LPH to generate mature MSH, resulting in the generation of MSH--endorphin fusion protein (13). This fusion protein bound to the human MC4R with an affinity similar to its natural ligands but had a markedly reduced ability to activate the receptor (13). Likewise, Krude et al. (36) also reported several missense heterozygous and homozygous POMC mutations in the coding sequence (Tyr5Cys) or regulatory and processing sites of the MSH gene. These missense mutations could generate nonfunctional or dominant-negative mutant MSH peptides that interfere with MC4R activity, resulting in obesity in human patients.
Early SAR studies to improve selectivity and potency of melanocortin peptides focused mostly on MSH analogs (16, 17, 18, 37). In addition, Peng et al. (38) examined the role of C-terminal peptide residues of MSH and MSH in imparting selectivity of the peptides toward MCRs. They concluded that proline12 (numbering with respect to MSH) played an important role for binding and activity at the MC1R but not at MC3R. More recently Kavarana et al. (39) discovered that another cyclic MSH peptide analog, (O)C-CH2-CH2-c[His6-D-Phe7-Arg8-Trp9-Lys10]amide, was a selective and potent agonist for human MC4R. Furthermore, Balse-Schnivasan et al. (40) synthesized novel MSH/MSH hybrid analogs that led to potent and selective ligands to human MC3R and MC4R. Finally, Oosterom et al. (41, 42) reported that a major determinant for selective receptor interaction is the conformational presentation of the core sequence in melanocortin peptides directed to the receptor binding pockets of the MC3R and MC4R.
Holder and Haskell-Luevano (43) recently published a comprehensive review covering 30 yr of SAR studies on melanocortin ligands. However, almost all previous SAR studies were focused on MSH analogs. Because we believe that MSH(5–22) is an important endogenous ligand for MC4R and MC3R, we decided to perform our SAR based on the MSH(5–22) structure. From a relatively focused SAR study on MSH(5–22) described in this report, we discovered that the unique Arg6 in MSH (5–22) (or Arg2 in peptide 6 as an equivalent residue) plays a crucial role in maintaining the ideal conformation of the cyclic peptide so that it can interact potently and selectively with the binding sites of MC4R and MC3R. We believe that further development of the potent and selective cyclic MSH-derived peptides such as peptide 6 [Ac-YRcyclo(CEHdFRWC)amid]e may lead to more novel and superior MC4R peptide agonists, useful for treating obesity or metabolic syndrome.
One major focus of this investigation was to study the acute and chronic effects of the selective MC4R/MC3R agonist (peptide 6) on energy balance (food intake, fat use) and understand its mechanism of action. To achieve this goal, we administered peptide 6 either centrally (ICV) or peripherally (sc) to rats and found that peptide 6 reduced food intake and increased fat use in rats, irrespective of their route of administration or the fat contents in diets. In addition, sc administration of a high dose (30 μmol/kg) of peptide 6 to mice reduced food intake and increased fat use only in WT but not MC4R-deficient mice. The data suggest the peptide 6 reduced energy balance in rodents, mostly through its specific effects on MC4R that is expressed almost exclusively in brain and hypothalamus. These acute biological data were also consistent with the data obtained from the chronicle administration of the peptide dosed peripherally in male DIO rats (Figs. 1 and 2). In addition, peptide 6 was also found to cause the stretching behaviors (Emmerson, P., et al., personal communication) similar to what was observed for MTII, an MC4R peptide agonist. The combined data strongly suggest an important role of central MC4R mechanism of action for the biological activity of peptide 6.
We did not perform a conditioned taste aversion test with peptide 6. However, we do not believe peptide 6 induced aversion behavior because direct or indirect activation of MC4R with central administration of MTII, MSH, or leptin does not cause taste aversion in rats at the dose that caused anorectic effects (44). Furthermore, peptide 6 caused no changes in locomotor activity or total EEs in treated rats (Fig. 1, D and E), suggesting no toxic effects of the compound. Finally, peptide 6 reduced food intake and fat use specifically on WT mice but not MC4R knockout mice, suggesting the effects on body weight and food intake were not mediated through nonspecific aversive or toxic effects of the compound but through the specific effects on MC4R.
Peptide 6 has also been observed to have extended half-life (>>2 h) in rat and human sera and have greater than 80% bioavailability in rats and dogs when administered sc (data not shown). Our data suggest that the potent in vivo activity of peptide 6 when administered peripherally is due to not only its intrinsic high efficacy for MC4R but also its great stability and bioavailability of the peptide in the treated animals.
To conclude, we discovered, through a focused SAR study, a novel and selective MC4R and MC3R peptide agonist, peptide 6. This synthetic cyclic peptide administered either centrally or peripherally can potently reduce energy balance, increase fat use, and decrease weight gain in acute and chronic metabolic studies in rats. This antiobesity effect by peptide 6 was manifested only in WT but not MC4R-deficient mice, indicating that the antiobesity effect was attributed mostly through the MC4R agonist activity of the peptide.
Acknowledgments
The authors are indebted to Drs. Jose Caro and Paul Burn for encouragement in this study and Dr. Paul Emmerson, Dr. Bob Gadski, and Mr. Steve Kahl for helpful suggestions. The authors are grateful to the technical support of Mr. Dave Flora, Ms. Xiaoying Gao, and Mr. Tom P. O’Brien.
Footnotes
First Published Online September 15, 2005
Abbreviations: DIO, Diet-induced obese; EE, energy expenditure; ICV, intracerebroventricularly; Ki, affinity constant obtained from competitive binding assays; LPH, -lipotropin; MCR, melanocortin receptor; MTII, Ac-Nle-cyclo(DHdFRWK)amide; NDP, Nle4, D-Phe7; NMR, nuclear magnetic resonance; PC, prohormone convertase; POMC, proopiomelanocortin; RQ, respiratory quotient; SAR, structure-activity relationship; WT, wild type.
Accepted for publication September 9, 2005.
References
Pritchard LE, Turnbull AV, White A 2002 Pro-opiomelanocortin processing in the hypothalamus: impact on melanocortin signalling and obesity. J Endocrinol 172:411–421
Cone RD, Lu D, Koppula S, Vage DI, Klungland H, Boston B, Chen W, Orth DN, Pouton C, Kesterson RA 1996 The melanocortin receptors: agonists, antagonists, and the hormonal control of pigmentation. Rec Prog Horm Res 51:287–317; discussion 318
Bennett HP 1986 Biosynthetic fate of the amino-terminal fragment of pro-opiomelanocortin within the intermediate lobe of the mouse pituitary. Peptides 7:615–622
Adan RA, Gispen WH 2000 Melanocortins and the brain: from effects via receptors to drug targets. Eur J Pharmacol 405:13–24
O’Rahilly S, Farooqi IS, Yeo GS, Challis BG 2003 Minireview: human obesity-lessons from monogenic disorders. Endocrinology 144:3757–3764
Yeo GS, Lank EJ, Farooqi IS, Keogh J, Challis BG, O’Rahilly S 2003 Mutations in the human melanocortin-4 receptor gene associated with severe familial obesity disrupts receptor function through multiple molecular mechanisms. Hum Mol Genet 12:561–574
MacNeil DJ, Howard AD, Guan X, Fong TM, Nargund RP, Bednarek MA, Goulet MT, Weinberg DH, Strack AM, Marsh DJ, Chen HY, Shen CP, Chen AS, Rosenblum CI, MacNeil T, Tota M, MacIntyre ED, Van der Ploeg LH 2002 The role of melanocortins in body weight regulation: opportunities for the treatment of obesity. Eur J Pharmacol [Erratum (2002) 450:93–109] 440:141–157
Roselli-Rehfuss L, Mountjoy KG, Robbins LS, Mortrud MT, Low MJ, Tatro JB, Entwistle ML, Simerly RB, Cone RD 1993 Identification of a receptor for melanotropin and other proopiomelanocortin peptides in the hypothalamus and limbic system. Proc Natl Acad Sci USA 90:8856–8860
Bray GA 1993 The nutrient balance hypothesis: peptides, sympathetic activity, and food intake. Ann NY Acad Sci 676:223–241
Harrold JA, Widdowson PS, Williams G 2003 MSH: a functional ligand that regulated energy homeostasis via hypothalamic MC4-R Peptides 24:397–405
Harrold JA, Williams G, Widdowson PS 1999 Changes in hypothalamic agouti-related protein (AGRP), but not MSH or pro-opiomelanocortin concentrations in dietary-obese and food-restricted rats. Biochem Biophys Res Commun 258:574–577
Hsiung HM, Smiley DL, Zhang X-Y, Zhang L, Yan LZ, Craft L, Heiman ML, Smith DP 2005 Potent peptide agonists for human melanocortin 3/4-receptors derived from enzymatic cleavages of human MSH(5–22) by dipeptidyl peptidase I and dipeptidyl peptidase IV. Peptides 26:1988–1996
Challis BG, Pritchard LE, Creemers JW, Delplanque J, Keogh JM, Luan J, Wareham NJ, Yeo GS, Bhattacharyya S, Froguel P, White A, Farooqi IS, O’Rahilly S 2002 A missense mutation disrupting a dibasic prohormone processing site in pro-opiomelanocortin (POMC) increases susceptibility to early-onset obesity through a novel molecular mechanism. Hum Mol Genet 11:1997–2004
Bertagna X, Lenne F, Comar D, Massias JF, Wajcman H, Baudin V, Luton JP, Girard F 1986 Human -melanocyte-stimulating hormone revisited. Proc Natl Acad Sci USA 83:9719–9723
Krude H, Biebermann H, Gruters A 2003 Mutations in the human proopiomelanocortin gene. Ann NY Acad Sci 994:233–239
Sawyer TK, Hadley ME, Hruby VJ, Castrucci AMD, Staples DJ, Farah J, O’Donohue TL 1988 -Melanocyte-stimulating hormone structure-activity studies: comparative analysis of melanotropic and CNS bioactivities. Synapse 2:288–292
Sawyer TK, Hruby VJ, Darman PS, Hadley ME 1982 [half-Cys4,half-Cys10]--melanocyte-stimulating hormone: a cyclic -melanotropin exhibiting superagonist biological activity. Proc Natl Acad Sci USA 79:1751–1755
Hruby VJ, Lu D, Sharma SD, Castrucci AL, Kesterson RA, al-Obeidi FA, Hadley ME, Cone RD 1995 Cyclic lactam -melanotropin analogues of Ac-Nle4-cyclo[Asp5, D-Phe7,Lys10] -melanocyte-stimulating hormone-(4–10)-NH2 with bulky aromatic amino acids at position 7 show high antagonist potency and selectivity at specific melanocortin receptors. J Med Chem 38:3454–3461
Merrifield RB 1963 Solid phase peptide synthesis. I. The synthesis of a tetrapeptide. J Am Chem Soc 85:2149–2154
Wang SS 1973 p-Alkoxybenzyl alcohol resin and p-alkoxybenzyloxycarbonylhydrazide resin for solid phase synthesis of protected peptide fragments. J Am Chem Soc 95:1328–1333
Rink H 1987 Solid-phase synthesis of protected peptide fragments using a trialkoxy-diphenyl-methylester resin. Tetrahedron Lett 28:3787–3790
Berg DT, Grinnell BW, McClure DB 1993 High-level expression of secreted proteins from cells adapted to serum-free suspension culture. Biotechniques 14:972–978
Chen Y, Heiman ML 2000 Chronic leptin administration promotes lipid utilization until fat mass is greatly reduced and preserves lean mass of normal female rats. Regul Pept 92:113–119
Flatt J 1991 Assessment of daily and cumulative carbohydrate and fat balances in mice. J. Nutr Biochem 2:193–202
Chen Y, Hu C, Hsu CK, Zhang Q, Bi C, Asnicar M, Hsiung HM, Fox N, Slieker LJ, Yang DD, Heiman ML, Shi Y 2002 Targeted disruption of the melanin-concentrating hormone receptor-1 results in hyperphagia and resistance to diet-induced obesity. Endocrinology 143:2469–2477
Liu J, Na S, Glasebrook A, Fox N, Solenberg PJ, Zhang Q, Song HY, Yang DD 2001 Enhanced CD4+ T cell proliferation and Th2 cytokine production in DR6-deficient mice. Immunity 15:23–34
Tinsley FC, Taicher GZ, Heiman ML 2004 Evaluation of a quantitative magnetic resonance method for mouse whole body composition analysis. Obes Res 12:150–160
Sawyer TK, Sanfilippo PJ, Hruby VJ, Engel MH, Heward CB, Burnett JB, Hadley ME 1980 4-Norleucine, 7-D-phenylalanine--melanocyte-stimulating hormone: a highly potent -melanotropin with ultralong biological activity. Proc Natl Acad Sci USA 77:5754–5758
Cody WL, Mahoney M, Knittel JJ, Hruby VJ, Castrucci AM, Hadley ME 1985 Cyclic melanotropins. 9. 7-D-phenylalanine analogues of the active-site sequence. J Med Chem 28:583–588
Hirsch MD, O’Donohue TL, Wilson R, Sawyer TK, Hruby VJ, Hadley ME, Cody WL, Knittel JJ, Crawley JN 1984 Structural and conformational modifications of MSH/ACTH4–10 provide melanotropin analogues with highly potent behavioral activities. Peptides 5:1197–1201
Ying J, Kover KE, Gu X, Han G, Trivedi DB, Kavarana MJ, Hruby VJ 2003 Solution structures of cyclic melanocortin agonists and antagonists by NMR. Biopolymers 71:696–716
Holder JR, Haskell-Luevano C 2003 Melanocortin tetrapeptides modified at the N-terminus, His, Phe, Arg, and Trp positions. Ann NY Acad Sci 994:36–48
Holder JR, Bauzo RM, Xiang Z, Scott J, Haskell-Luevano C 2003 Design and pharmacology of peptoids and peptide-peptoid hybrids based on the melanocortin agonists core tetrapeptide sequence. Bioorg Med Chem Lett 13:4505–4509
Bertagna X 1994 Proopiomelanocortin-derived peptides. Endocrinol Metab Clin North Am 23:467–485
Raffin-Sanson ML, de Keyzer Y, Bertagna X 2003 Proopiomelanocortin, a polypeptide precursor with multiple functions: from physiology to pathological conditions. Eur J Endocrinol 149:79–90
Krude H, Biebermann H, Schnabel D, Tansek MZ, Theunissen P, Mullis PE, Gruters A 2003 Obesity due to proopiomelanocortin deficiency: three new cases and treatment trials with thyroid hormone and ACTH4–10. J Clin Endocrinol Metab 88:4633–4640
Danho W, Swistok J, Cheung AW, Chu XJ, Wang Y, Chen L, Bartkovitz D, Gore V, Qi L, Fry D, Greeley D, Sun H, Guenot J, Franco L, Kurylko G, Rumennik L, Yagaloff K 2001 Highly selective cyclic peptides for human melanocortin-4 receptor (MC4-R): design, synthesis, bioactive conformation, and pharmacological evaluation as anti-obesity agent. Peptides: the wave of the future. Proc Second International and the Seventeenth American Peptide Symposium, San Diego, CA, 2001, pp 701–703
Peng PJ, Sahm UG, Doherty RV, Kinsman RG, Moss SH, Pouton CW 1997 Binding and biological activity of C-terminally modified melanocortin peptides: a comparison between their actions at rodent MC1 and MC3 receptors. Peptides 18:1001–1008
Kavarana MJ, Trivedi D, Cai M, Ying J, Hammer M, Cabello C, Grieco P, Han G, Hruby VJ 2002 Novel cyclic templates of MSH give highly selective and potent antagonists/agonists for human melanocortin-3/4 receptors. J Med Chem 45:2644–2650
Balse-Srinivasan P, Grieco P, Cai M, Trivedi D, Hruby VJ 2003 Structure-activity relationships of MSH analogues at the human melanocortin MC3, MC4, and MC5 receptors. Discovery of highly selective hMC3R, hMC4R, and hMC5R analogues. J Med Chem 46:4965–4973
Oosterom J, Nijenhuis WA, Schaaper WM, Slootstra J, Meloen RH, Gispen WH, Burbach JP, Adan RA 1999 Conformation of the core sequence in melanocortin peptides directs selectivity for the melanocortin MC3 and MC4 receptors. J Biol Chem 274:16853–16860
Oosterom J, Garner KM, den Dekker WK, Nijenhuis WA, Gispen WH, Burbach JP, Barsh GS, Adan RA 2001 Common requirements for melanocortin-4 receptor selectivity of structurally unrelated melanocortin agonist and endogenous antagonist, agouti protein. J Biol Chem 276:931–936
Holder JR, Haskell-Luevano C 2004 Melanocortin ligands: 30 years of structure-activity relationship (SAR) studies. Med Res Rev 24:325–356
Wirth MM, Olszewski PK, Yu C, Levine AS, Giraudo SQ 2001 Paraventricular hypothalamic -melanocyte-stimulating hormone and MTII reduce feeding without causing aversive effects. Peptides 22:129–134(Hansen M. Hsiung, JeAnne Hertel, Xing-yu)
Lead Optimization Biology (S.H.), Eli Lilly & Co., Indianapolis, Indiana 46285
Abstract
MSH has generally been accepted as the endogenous ligand for melanocortin 4 receptor (MC4R), which plays a major role in energy homeostasis. Targeting MC4R to develop antiobesity agents, many investigators have performed a structure-activity relationship (SAR) studies based on MSH structure. In this report, we performed a SAR study using human MSH (5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ) (DEGPYRMEHFRWGSPPKD, peptide 1) as a lead sequence to develop potent and selective agonists for MC4R and MC3R. The SAR study was begun with a truncation of N terminus of MSH (5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 ) together with acetylation of the N terminus and amidation of the C terminus of the peptide. Introduction of a cyclic disulfide constrain and replacement of L-Phe with D-Phe afforded a super potent agonist (peptide 5). Furthermore truncation at the C terminus generated a small and potent MC4R and MC3R agonist (Ac-YRcyclo[CEHdFRWC]amide, peptide 6), which exhibited no MC5R and greatly reduced MC1R activity. Molecular modeling of Ac-YRcyclo[CEHdFRWC]amide (peptide 6) revealed that Arg2 in the peptide formed a salt bridge with Glu4. Subcutaneous or intracerebroventricular administration of peptide 6 in rats showed potent in vivo efficacy as evidenced by its effects in reducing energy balance, increasing fat use, and decreasing weight gain in both acute and chronic rat metabolic studies. Furthermore, the antiobesity effect by peptide 6 was manifested only in wild-type but not MC4R-deficient mice, indicating that antiobesity effects of the peptide were attributed largely through MC4R but not MC3R agonist activity of the peptide.
Introduction
THE PROOPIOMELANOCORTIN (POMC) gene expression in tissues produces a 36-kDa preprohormone, which is cleaved by a signal peptidase to a 32-kDa prohormone, POMC (1). POMC protein (32 kDa) is further processed in a tissue-specific manner by various processing enzymes to generate multiple and diverse peptide hormones including MSH, MSH, MSH, ACTH, -lipotropin (LPH), and -endorphin (1). Specifically, proPOMC protein is processed first by prohormone convertase (PC)1 in tissues to generate proACTH (22 kDa) and LPH (10 kDa). ProACTH is further processed to form N-terminal POMC peptide and ACTH (4 kDa). ACTH is then processed further by PC2, carboxypeptidase E, peptidyl -amidating monooxygenase, and N-acetyltransferase to generate mature MSH and des-acetyl-MSH (1, 2). In most mammalian species, LPH is processed by PC2 to generate MSH and -endorphin. However, due to the lack of dibasic residues at the MSH processing site in the rodent LPH sequence, mature MSH cannot be generated in rodents (1, 3). Furthermore, because PC2 is not present in the human pituitary gland, the main POMC-derived peptides in human pituitary are N-terminal POMC peptide, ACTH, and LPH whereas MSH and MSH are produced in hypothalamus, brain, and some peripheral tissues (1).
Diverse lines of evidence, including genetic and pharmacological data obtained in rodents and humans, suggest that the melanocortin 3 receptor (MC3R) and MC4R play important roles in the regulation of energy homeostasis in animals (4). In fact, MC4R mutation is the most prevalent single gene mutation that caused human obesity (5, 6). The human and rodent genetic data strongly support that MC4R is the most validated target for obesity and the POMC-derived peptides are native ligands for MC4R and MC3R. Among the POMC-derived peptides, MSH was long presumed to be the primary ligand for MC4R (7), and MSH was regarded as the primary ligand for MC3R (8). On the other hand, MSH’s role with respect to the MCRs is not fully understood, and it is not considered to be a major or important ligand for any of these receptors.
Although MSH was considered to be the primary peptide ligand for MC4R and the peptide was shown to decrease feeding in rodents when injected intracerebroventricularly (ICV) (9), MSH binds to cloned human MC4R or rat hypothalamic homogenate with only moderate affinity [inhibitory constant (Ki) = 324 and 22.5 nM, respectively] (10). Additionally, Harrold et al. (11) reported that MSH concentration in the hypothalamus was not changed, whereas the concentration of an endogenous MC4R and MC3R antagonist, agouti-related protein, was altered by changes in the nutritional status of the animals. These data suggest that, in addition to MSH, perhaps other peptide or protein ligands may also play a role in regulating energy homeostasis by interacting with MC4R and MC3R in hypothalamus.
Several recent reports suggested that MSH may serve as a particularly important endogenous ligand for MC4R and MC3R in the regulation of energy homeostasis. First, MSH bound cloned human MC3R and MC4R with equal or higher affinity when compared with MSH (10, 11, 12, 13). Second, MSH is produced at relatively high concentrations in the hypothalamus (1, 10, 14), a key region for regulating energy homeostasis. Finally, some POMC mutations associated with human obesity are localized in the coding region or processing sites of the human MSH gene (13, 15). These mutated POMC genes would generate a MSH--endorphin fusion protein and truncated MSH peptides. The MSH--endorphin fusion protein was shown to be a potent antagonist to MC4R. Therefore, the mutant protein can disrupt the function of the native ligand (MSH) with MC4R by a dominant-negative effect (13). Alternatively, some human mutations that produced truncated MSH or no MSH also exhibited obese phenotype (15). These combined data suggest that the MSH peptide could play a physiologically important role in regulating energy homeostasis.
Early melanocortin peptide structure-activity relationships (SARs) studies focused mostly on MSH and its analogs (16, 17, 18). As a result, several potent but nonselective melanocortin peptides, including (Nle4, D-Phe7) (NDP)-MSH, MTII [Ac-Nle-cyclo (DHdFRWK)amide] and SHU9119 [Ac-Nle-cyclo(DHdNalRWK)amide] were developed as important tools to study MCR pharmacology. The underlined and italicized amino acid sequences denote the core peptide sequences for MCR binding. In this report, we describe a focused SAR study using human MSH(5–22) as a lead sequence. Our SAR study, based on the MSH structure, generated several potent and selective MC4R and MC3R peptide agonists. Additionally, we observed that a potent and selective MC4R and MC3R peptide agonist, Ac-YRcyclo[CEHdFRWC]amide (peptide 6 in Table 1) reduced food intake and body weight in diet-induced obese (DIO) rats and normal chow-fed rats. Finally, we demonstrate that the effect of the peptide 6 was mediated through its MC4R activity as shown by its lack of activity in MC4R-deficient mice.
Materials and Methods
Peptide synthesis
All peptides were synthesized by solid-phase methods (19). The solid supports were either preloaded Wang resin for C-terminal acid or MBHA Rink resin for C-terminal peptide amide (20, 21). The primary peptide chain was assembled using an ABI 433A synthesizer (PE Applied Biosystems Inc., Foster City, CA). Side-chain protection of amino acids were compatible to standard Fmoc chemistry, as shown below: Arg(Pbf), Asp(OtBu), Cys(Trt), Glu(OtBu), Gln(Trt), His(Trt), Lys(Boc), Ser(tBu), Thr(tBu), Trp(Boc), Tyr(tBu), or otherwise specified. Other common and uncommon amino acids were used without side-chain protection. Single coupling of each residue was performed using dicyclohexylcarbodiimide/hydroxybenztriazole activation protocol, with a 4-fold excess of each amino acid and 1.5 h coupling time. Acetylation of the -amino group after the chain assembly was normally carried out off-line with 5 eq acetic anhydride, 10 eq N,N-diisopropylethyl amine in dry N-N-dimethylformamide for 1 h at room temperature. The finished peptide was simultaneously deprotected and cleaved from the resin using a cocktail of trifluoroacetic acid/H2O/triisopropyl silane (Tis)/1,2-ethanedithiol (95/2/1/2, vol/vol), or trifluoroacetic acid/H2O/Tis/anisole (92/2/4/2, vol/vol) for 2 h at room temperature. The solvents were then evaporated under vacuum, and the residue was washed three times with cold ether to remove the scavengers. The residue was redissolved in aqueous acetonitrile and purified on reverse phase-HPLC. In case of peptides with disulfide bond, the crude product was dissolved in dimethylsulfoxide and then diluted with 0.2 M ammonium acetate buffer (pH 7) to a final concentration of 20% of dimethylsulfoxide. The solution was agitated at room temperature overnight to form the disulfide bond. Then the solution was diluted with two volumes of water and purified by a preparative RP-HPLC. Depending on the amount of peptide to be purified, different size of the reverse-phased column (Vydac or Zorbax C18, diameter ranges from 0.9 to 5.0 cm) was chosen accordingly. The fractions containing the desired product were pooled and lyophilized. Further characterization of the final product was performed using analytical HPLC and mass and UV spectrometry.
MCR binding assays
We performed radioligand-binding assay by using 125I-NDP-MSH (sp act. 2000 Ci/mmol; Amersham, Piscataway, NJ) as the radioactive ligand. Cell membranes were prepared from the human embryonic kidney 293 cells stably transfected with cloned human MCRs. Cells were grown as adherent monolayers in roller bottle cultures at 37 C and 5% CO2/air atmosphere in a 3:1 mixture of DMEM and Ham’s F12 containing 25 mM l-glucose, 100 U/ml penicillin G, 100 μg/ml streptomycin, 250 ng/ml amphoterin B, 300 μg/ml gentisin, and supplemented with 5% fetal bovine serum. For large-scale production, monolayer cells are adapted to suspension culture (22) and are grown in either spinner or shaker flasks (37 C and 7.5% CO2/air overlay) in a modified DME/F12 medium containing 0.1 mM CaCl2, 2% equine serum, and 100 μg/ml sodium heparin to prevent cell-cell aggregation. Cells are harvested by centrifugation, washed in PBS, and pellets stored frozen at –80 C.
For the preparation of membranes, frozen cell pellets were resuspended in 10 volumes of membrane prep buffer (10 ml buffer per gram of cell paste) consisting of 50 mM Tris (pH 7.5) at 4 C, 250 mM sucrose, 1 mM MgCl2, Complete EDTA-free protease inhibitor tablet (Roche Applied Science, Indianapolis, IN), and 24 μg/ml DNase I (Sigma, St. Louis, MO). The cells were homogenized with a motor-driven Teflon-glass Dounce tissue homogenizer (Wheaton Science Products, Millville, NJ) using 20 strokes, followed by centrifugation at 38,000 x g at 4 C for 40 min. The pellets were resuspended in membrane prep buffer at a concentration of 2.5–7.5 mg/ml, aliquoted, quick frozen in liquid nitrogen, and stored at –80 C.
Standard competitive receptor/ligand binding experiments were performed in 96-well plate formats, and the assay mixtures consisted of serial dilutions of test compounds (10 μM to 100 pM) or those of unlabeled NDP-MSH (100 nM to 1 pM) in binding buffer [25 mM HEPES (H 7.5); 10 mM CaCl2; 0.3% BSA]. The incubation mixture also contained 0.5–5.0 μg membrane proteins, 100 pM 125I-NDP-MSH, and 0.25 mg of wheat germ agglutinin SPA beads. The mixture solutions in 96-well plates were then agitated briefly on a plate shaker and incubated for 10 h at room temperature. The radioactivity bound to the receptor was quantified in a Trilux microplate scintillation counter (PerkinElmer Life Sciences, Norwalk, CT). Nonlinear regression analysis of competitive binding assay data using a four-parameter logistic fit yielded IC50 values, which were converted to affinity constants obtained from competitive binding assays (Ki values). These conversions were performed by using the Cheng-Prusoff equation, Ki = IC50/(1 + D/Kd), where D is the concentration of radioligand and Kd is the equilibrium dissociation constant determined from saturation binding analysis.
In vitro MCR functional assays
Functional activity was determined using a standard cAMP assay with serial dilutions of test compound (10 μM to 0.1 nM) or the control agonist NDP-MSH (100 nM to 1 pM). The human embryonic kidney 293 cells stably transfected with the human MC1R, MC2R, MC3R or MC4R were grown in DMEM containing 10% fetal bovine serum and 1% antibiotic/antimycotic solution. On the day of the assay, the cells were dislodged with enzyme-free cell dissociation solution and resuspended in cell buffer (Hanks’ balanced salt solution without phenol red Hanks’ balanced salt solution-092, 0.1% BSA, 10 mM HEPES) at 1 x 106 cells/ml. Cell suspension (40 μl) was added to positron emission tomography 96-well plates containing 20 μl of diluted compound or control agonist. Plates were incubated at 37 C for 20 min, and the assay was stopped by the addition of 50 μl of quench buffer (50 mM Na acetate, 0.25% Triton X-100).
Per the manufacturer’s (Amersham) instructions, cAMP concentrations were determined by a scintillation proximity assay-based competition assay using 125I-cAMP (Amersham), goat anti-cAMP antibody (MP Biomedicals, Irvine, CA), and polyvinyl toluene antisheep antibody binding beads (Amersham). The assay buffer contained 50 mM sodium acetate and 0.1% BSA. A mixture containing scintillation proximity assay beads (1 mg/ml), antibody (0.65%), and radioligand (61 pM) was prepared in assay buffer, and 100 μl of this solution were added to each well of the 96-well assay plate to yield a final volume of 210 μl. After a 12-h incubation, the plates were counted in a Trilux microplate scintillation counter (PE Life Sciences). The data were converted to picomoles of cAMP using a standard curve obtained from the same assay performed with varying concentrations of unlabeled cAMP. The data were analyzed using a four-parameter logistic nonlinear regression to generate agonist potencies (EC50) and percentages of efficacy relative to the maximum stimulation obtained with NDP-MSH.
Animal and maintenance
All animal experiments were conducted in accordance with principles and procedures outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals. In addition, Lilly Research Laboratories Animal Care and Use Committee approved all animal protocols used in these studies. Male Long-Evans rats were weaned and maintained on a diet comprised of 40% fat, 40% carbohydrate, and 20% protein (Teklad 95217; Harlan, Indianapolis, IN) and maintained at Harlan Sprague Dawley (Indianapolis, IN) until 90 d of age when they were used. Rats (500 g) were acclimated (single housed) for at least 1 wk in a temperature- (25 C) and light-controlled (12-h light, 12-h dark; lights on at 2200 h) environment. Water and food were available ad libitum.
ICV cannulation and administration
Male Long-Evans rats (Harlan) approximately 275 g (60 d old) on a normal chow diet (Purina 5008) were implanted with stainless steel cannulae guides (5 mm length, 26 gauge; Plastics One, Roanoke, VA) stereotaxically implanted in the lateral ventricle anteroposteriority: 0.8 mm caudal to bregma; and lateral: 1.5 mm from midline suture. Animals were handled and weighed daily and were placed in calorimeter chambers on the day previous to the test day to minimize stress. On the test day, animals were injected between 1300 and 1400 h with 5 μl of the peptide 6 (0.187 nmol) or saline, and indirect calorimetry was performed for 22–24 h after injection as described below.
Indirect calorimetry
Indirect calorimetry experiments were performed following a published procedure (23). Specifically, 24-h caloric expenditure and respiratory quotient (RQ) were measured by using an open circuit system (Oxymax; Columbus Instruments, Columbus, OH). RQ was measured as the ratio of volume of CO2 produced to the volume of O2 consumed. Caloric expenditure was calculated as the product of calorific value of oxygen and volume of O2 consumed per kilogram of body weight, where calorific value of oxygen = 3.815 + 1.232 (RQ). Total calories expended were calculated to determine daily fuel use. To calculate the proportion of protein, fat and carbohydrate that used during the 24-h period, we used Flatt’s proposal and formula as other derived constants (24). Locomotor activity was measured by counting number of new beam breaks (ambulatory) and same beam breaks (fine movement) during each period of 24 h in the calorimeter. Comparisons between groups of data (mean ± SEM) were analyzed using SigmaStat (SPSS Inc., Chicago, IL) ANOVA using Student-Newman-Keuls (Student-Newman-Keuls) test for multiple comparisons.
MC4R-deficient mice
Generation of MC4R-deficient mice is described as follows: murine genomic DNA clones corresponding to the MC4R locus were isolated from a mouse strain 129/SvJ (Fix II phage library, Stratagene Inc., La Jolla, CA) using a probe generated by PCR amplification of murine MC4R coding sequences. A gene-targeting backbone vector pGT-NN-tk was constructed as described (25). The targeting vector pKO-MC4R contained a 3.2-kb HindIII-NcoI fragment obtained from the 5' end of a MC4R genomic clone and a 1.4-kb HindIII-SacI fragment obtained from the 3' end of a MC4R genomic clone using pGT-NN-tk as a backbone. The gene targeting construct replaced the entire MC4R gene (an internal 1.5-kb NcoI-HindIII genomic fragment) with the neomycin resistance cassette. Electroporation of the linearized pKO-MC4R plasmid DNA into R1 embryonic stem (ES) cells, selection of G418 and ganciclovir-resistant clones, and generation of chimeras were performed as described (26). Male chimeras were bred to female C57BL/6 mice to generate F1 heterozygous mice. Germline transmission of MC4R mutation was confirmed by Southern analysis of tail DNA. The 10th-generation mice of C57BL/6 background heterozygous for the disrupted MC4R allele were bred to generate MC4R-null and wild-type (WT) littermates (Taconic, Hudson, NY). Heterozygous MC4R(+/–) mice were interbred to generate MC4R(–/–) and WT littermates at Taconic Farm. MC4R(–/–) and WT littermates were weaned at 3 wk onto normal chow diet (NIH-31; Ziegler, Gardners, PA) and shipped to Lilly Research Laboratories at 8 wk. Upon arrival the mice were housed individually and permitted to eat (TD 95217; Teklad, Madison, WI) and drink ad libitum. Animals were acclimated for at least 2 wk in a temperature- (25 C) and light-controlled (12-h light, 12-h dark; lights on at 2200 h) environment.
Acute metabolic study
Male Long-Evans rats.
Rats (500 g) were acclimated (single housed) for at least 1 wk in a temperature- (25 C) and light-controlled (12-h light, 12-h dark; lights on at 2200 h) environment. Water and food were available ad libitum. Groups of four rats each were fasted for 4 h and treated once between 1300 and 1400 h by sc injection with vehicle (saline) or peptide 6 (0.3, 1.5, and 3 μmol/kg). The rats were placed in sealed chambers to determine acute changes in energy expenditure (EE) and respiratory exchange ratio by indirect calorimetry (Oxymax; Columbus Instruments International). Rats were permitted to eat a high-fat diet containing 20% protein, 40% fat, and 40% carbohydrate (TD 95217; Teklad) and drink ad libitum throughout the study.
Male MC4R-deficient mice.
Groups of WT (n = 4) and MC4R-null mice (n = 4) were treated once between 1000 and 1100 h by sc injection with vehicle (saline) or peptide 6 (0.3, 3, or 30 μmol/kg). The mice were placed in sealed chambers to determine acute changes in EE and respiratory exchange ratio by indirect calorimetry (Oxymax; Columbus Instruments). Mice were permitted to eat a high-fat diet containing 20% protein, 40% fat, and 40% carbohydrate (TD 95217; Teklad) and drink ad libitum throughout the study.
Chronic metabolic activity study
Rats were randomized to three groups and each group was comprised of five animals. The animals were housed individually throughout the study and permitted to eat (TD 95217; Teklad) and drink ad libitum throughout the study.
Change in total body adipose and lean mass during treatment was measured by nuclear magnetic resonance (NMR) (Echo Medical Systems; Houston, TX) (27) before treatment and on d 15 of treatment. The wide-line NMR signal was calibrated using chicken breast muscle (fat and skin removed) for lean mass and canola oil for fat mass (27). Initial measurement of body composition by NMR was performed 1 d before initiation of treatment and was measured again on d 15. All treatments began on d 1 at 0900 h and continued for 14 d. Groups of five rats each were treated daily by sc injection with vehicle (saline, 1 ml/kg) or peptide 6 (0.3 and 3 μmol/kg). The rats were placed in sealed chambers to determine acute changes in EE and RQ by indirect calorimetry (Oxymax) on d 1, 7, and 14. Animals were then returned to their home cages after each calorimetry session, and daily treatment continued. Body weight and food intake were measured daily at 0900 h.
Serum hormone measurements (insulin, ghrelin, and leptin) and blood chemistry (glucose and corticosterone) for the rats treated with vehicle (saline, 1 ml/kg) or peptide 6 (0.3 and 3 μmol/kg) were performed by Linco Diagnostic Services, Inc. (St. Charles, MO). The rats (five rats per group) were treated in the same manner as described above except the treatment period was shortened to 10 d and the rats were then killed for serum collection.
Statistical analysis
All data are shown as the mean SEM, and statistical comparisons were made using ANOVA followed by Student-Newman-Keuls test when significant (P < 0.05) interaction was found and more than two groups were being compared.
Model structure of Ac-YRcyclo[CEHdFRWC]amide (peptide 6)
Model structure of peptide 6 (Table 1) was built using Quanta98 (Accelrys, San Diego, CA). Energies, including contributions from Van der Waals and electrostatic terms, were minimized in vacuum. Turning on and off explicit hydrogen bonds had no significant impact on the model structure because the cyclic nature of this molecule greatly limits the freedom of side chain placement.
Results
SAR studies
Our previous studies with synthetic melanocortin peptides (12) demonstrated that human MSH(5–22) (peptide 1) and particularly its diaminopeptidase (I and IV) cleaved products are potent but nonselective MC4R agonists. Aiming to develop a more selective agonist for MC4R and MC3R, we performed a further SAR study (Table 1) using human MSH(5–22) DEGPYRMEHFRWGSPPKD (peptide 1) as a lead sequence. Human MSH(5–22) (peptide 1) is a moderately potent agonist with Ki = 23 nM and EC50 = 9.3 nM for MC4R (Table 1). By removing two and four amino acid residues, respectively, from the N terminus of human MSH(5–22), we obtained two nonselective, yet more potent, MC4R peptide agonists, MSH(7–22) (peptide 2, Ki = 4.6 nM) and MSH(9–22) (peptide 3, Ki = 5.7 nM) (Table 1). The binding affinity and functional potency of these two peptides toward MC4R and MC3R were severalfold greater than their predecessor peptide, MSH(5–22).
To increase the stability of the peptide against serum proteases, we further modified the structure of the peptide MSH(9–22) by capping the N terminus with an acetyl group, converting the C terminus acid to an amide, and replacing L-Phe with D-Phe. With these changes, we obtained a super potent but nonselective peptide agonist, peptide 4 (Table 1) for MC1R, MC3R, MC4R and MC5R. Our result with the replacement of L-Phe by D-Phe residue in MSH(5–22) peptide was consistent with the previous literature results with the same modification of MSH-derived peptides (17, 28). This D-Phe replacement not only increases proteolytic resistance of the peptide but also enhances its binding affinity to MC4R by stabilizing the turn structure (17). Aiming to further increase selectivity of the MSH analogs, we altered the peptide structure by incorporating two cysteines in the molecule and generated a cyclic MSH(9–22) analog with a disulfide bond similar to what was previously reported in MSH-derived peptides (17). The resulting cyclic peptide 5 (Table 1) showed a more selective and potent agonist activity for MC4R with no detectable MC5R agonist activity (Ki for MC4R = 0.5 nM and Ki for MC5R > 500 nM). Additionally, removal of the last five residues (SPPKD) from the C terminus of peptide 5 afforded peptide 6 (Table 1), which was a highly potent MC4R agonist (Ki = 0.8 nM; EC50 = 0.6 nM) but showed a much reduced binding affinity toward MC1R (Ki = 17.8 nM). In fact, peptide 6 (MC1R, Ki = 17.8 nM) showed a 30-fold lower MC1R binding affinity than its predecessor peptide, 5 (MC1R, Ki = 0.6 nM).
The cyclic disulfide peptide structure appears to be crucial for its MC4R potency and selectivity. Disrupting the ring structure by replacing cysteines with -aminobutyric acid in peptide 6 resulted in linear peptide 9 that showed significantly compromised MC4R activity and selectivity (Table 1). The D-Phe residue in the center of the sequence appeared to be essential for its affinity toward MC4R. Replacing D-Phe in peptide 6 with L-Phe afforded peptide 10, which exhibited greatly reduced MC4R binding affinity (Ki = 30.5 nM), compared with peptide 6 (MC4R, Ki = 0.8 nM). Furthermore, cyclic peptide 10 also showed much weaker binding activity toward MC1R, MC3R, and MC5R (Table 1).
The SAR data (Table 1) also showed that peptide 6 appeared to reach an optimal MC4R activity or potency. Removal of tyrosine from the N terminus of peptide 6 resulted in peptide 7, which exhibited slightly more potent binding affinity and functional activity for MC4R (Ki = 0.4 nM, EC = 0.4 nM) and MC3R (Ki = 39.4 nM). The exocyclic basic amino acid residue (Arg) present in the MSH(5–22)-derived peptides such as peptides 6 and 7 appeared to be crucial for its potency and activity toward MC4R and MC3R. Additional removal of Arg residue from amino terminus of peptide 7 generated peptide 8 (Table 1), which had weaker binding affinity and functional activity for MC4R (Ki = 2.5 nM, EC50 = 3.3 nM) and MC3R (Ki = 478 nM) than peptides 6 or 7. This less potent MC4R peptide 8 has been previously reported to be an active melanotropin in frog or lizard skin bioassay (29) and a neuroactive substance causing grooming behavior in rats after ICV injections (30).
Acute energy balance changes in DIO rat models administered with peptide 6
Various dosages (0.3, 1.5, and 3 μmol/kg) of peptide 6 together with saline control were administered sc into the male Long-Evans rats maintained on a caloric-dense diet. Subcutaneous administration of peptide 6 (0.3, 1.5, and 3 μmol/kg) in rats reduced food intake to 57, 45, and 25% of the levels of the vehicle control rats treated with saline (Fig. 1A). It was also observed that those rats treated with peptide 6 showed significant increases in fat use [126, 157, and 157% of those of the vehicle controls (Fig. 1B)]. Together we demonstrated that the peptide-treated rats exhibited a negative energy balance in a dose-dependent manner. Treatment with peptide 6 also decreased RQ in DIO rats when compared with vehicle control (Fig. 1C). Additionally, we observed no significant changes in total EE or locomotor activity in the rats treated with the peptide when compared with those of saline-treated rats (Fig. 1, D and E).
Chronic energy balance changes in body weight and lean and fat mass of the DIO rats
Ac-YRcyclo[CEHdFRWC]amide (peptide 6, Table 1) was administered to male Long-Evans rats once daily for 14 d at two different dosages (0.3 and 3.0 μmol/kg). During this period, we measured the changes in body weight and body composition of the rats. Figure 2A shows that peptide 6 at a dose of 3 μmol/kg caused a significant decrease in body weight throughout the 2-wk testing period. Administration of the peptide at 10-fold lower dose (0.3 μmol/kg) also caused significant decrease in body weight at the end of the 2-wk testing period. Furthermore, we observed a gradual rebound in body weight after the administration of the peptide was stopped (Fig. 2A).
Additionally, Fig. 2B shows that rats treated with peptide 6 (3 μmol/kg) exhibited a significant reduction (>20 g) in fat mass with no change in lean mass when compared with saline vehicle controls. We also observed a trend of decreases in serum leptin or triglyceride concentration in the rats treated for 10 d with a high dose of peptide 6 (3 μmol/kg) when compared with those of vehicle controls. However, no changes of serum insulin, glucose, ghrelin, or corticosterone were observed in all treated rats (data not shown).
Acute ICV administration of peptide 6 in rats fed regular chow
To further test the central melanocortin activity of the peptide, we administered the peptide 6 ICV at a very low dose (0.187 nmol/rat) to rats fed on a regular chow. The results (Fig. 3) showed that ICV administration of peptide 6 at a low dose reduced food consumption to 68% (Fig. 3A) and increased fat use to 439% of those observed in vehicle controls (Fig. 3B) over a 24-h period of time. Indirect calorimetry also showed a significant reduction of RQ in the rats treated with peptide 6 when compared with the vehicle control rats (data not shown).
In vivo activity of peptide 6 in WT and MC4R-deficient mice
To further elucidate the mechanism of in vivo activity of peptide 6, we administered this peptide to the MC4R-deficient (MC4R–/–) mice and their sibling WT controls. The results (Fig. 4A) showed that sc administration of peptide 6 at a high dose (30 μmol/kg) to the MC4R-deficient mice showed no effects on food consumption over a 24 h period; however, the treatment of peptide 6 in WT sibling control mice caused a dramatic decrease in food intake (to 19% of WT vehicle control). Furthermore, the results (Fig. 4B) also showed the dose-dependent increases of fat use (158, 203, and 221% of the vehicle control) in WT mice treated with peptide 6. On the other hand, the peptide caused no statistically significant changes in fat use in the MC4R–/– mice (Fig. 4B). These results demonstrate that the major antiobesity effect of peptide 6 was mediated through an MC4R mechanism and not through MC3R activity of the peptide.
Model structure of peptide 6
The model structure of peptide 6 (Fig. 5; first residue as Tyr1) showed that exocyclic Tyr1 residue is well separated from the cyclic core structure that is generally believed to be the main recognizing element in the receptor binding (31). As expected from this model, peptide 7, which was resulted from removal of N-terminal Tyr1 from peptide 6, retains its potency for MC3R and MC4R. However, the second exocyclic residue in peptide 6, Arg2, was shown to interact with units within the core structure, i.e. the side chain of Arg2 forms a strong salt bridge with the side chain of Glu4 (Fig. 4). This ionic interaction may stabilize the main cyclic structure in addition to the stabilization afforded by D-Phe6. It was interesting to note that the side chain of Glu4 was pulled away from the cyclic core structure, presumably due to formation of this salt bridge. If bothTyr1 and Arg2 were removed, the resulting compound, i.e. Ac-cyclo[CEHdFRWC]amide (peptide 8, Table 1), would lose Arg2-Glu4 charge interactions and have flexible side chains. We speculated that the cyclic core structure with its two positive charges (His3 and Arg7) could be the key recognition elements for a negatively charged pocket in MCRs (32). Absence of Arg2 would not only potentially weaken ligand-receptor electrostatic interaction but also make binding entropically unfavorable. Indeed, the Ki values of peptide 8 for MC4R and MC3R were reduced by severalfold (MC4R Ki = 2.5 nM and MC3R Ki = 478 nM, respectively) as compared with those of peptide 6 (MC4R Ki = 0.8 nM, MC3R Ki = 56.4 nM). Thus, Arg2 may have two roles: it neutralizes the Glu4-negative charge so that the positive charges (His3 and Arg7) in the core domain can be recognized by the receptors through charge interaction, and it limits Glu4 side-chain freedom (reduces Glu4 entropy in unbound state) and stabilizes its conformation.
The role of D-Phe provided extra stability to the -turn for the cyclic peptide. Our model structure showed the side chain of His3 might have -stack-like interactions with the side chain of D-Phe6. Replacing D-Phe with L-Phe would be less likely to have such an effect. Furthermore, an electrostatic surface potential plot (not shown) generated from our model shows that Trp8 and D-Phe6 form a hydrophobic extrusion, whereas His3 and Arg7 on the other side have positive charges exposed. Our model supports a recent report (33) that both the hydrophilic surface and the hydrophobic surface are needed for tight binding to the melanocortin receptors.
Discussion
Central MC4R and MC3R are known to play pivotal roles in regulating energy homeostasis in animals. The endogenous ligand for MC4R was long presumed to be MSH, a 13-amino acid peptide hormone produced in hypothalamus, brain, pituitary, and many other peripheral tissues. A primary ligand for MC3R has been hypothesized to be MSH, which has shown selective activity toward MC3R (8). However, MSH such as human MSH(5–22) (DEGPYRMEHFRWGSPPKD, peptide 1, Table 1) isolated from human hypothalamus, was not considered to be an endogenous ligand for MC4R or MC3R, even though it has moderately potent agonist activities for MC3R and MC4R (Table 1).
Bertagna et al. (14) showed that human MSH(5–22) is a true endogenous peptide isolated from human hypothalamus, whereas previously reported 22-amino acid human MSH(1–22) peptide (AEKKDEGPYRMEHFRWGSPPKD) was generated artificially from acid extraction of the peptides from human pituitary. Recently Harrold et al. (10) and we (12) have also proposed that MSH in the hypothalamus is an endogenous MC4R ligand whose activity may play an important role in regulating energy homeostasis. Finally, MSH(1–22) is also a proteolytic fragment corresponding to residues 35–56 of human LPH (34, 35).
That MSH may play an important physiological role in human energy homeostasis has been further supported by human genetic studies (13, 36). Human heterozygous POMC gene mutation (Arg236Gly) causing obese phenotype has been reported at the junction of MSH--endorphin processing site of the gene (13). The patient can produce native MSH but is unable to process LPH to generate mature MSH, resulting in the generation of MSH--endorphin fusion protein (13). This fusion protein bound to the human MC4R with an affinity similar to its natural ligands but had a markedly reduced ability to activate the receptor (13). Likewise, Krude et al. (36) also reported several missense heterozygous and homozygous POMC mutations in the coding sequence (Tyr5Cys) or regulatory and processing sites of the MSH gene. These missense mutations could generate nonfunctional or dominant-negative mutant MSH peptides that interfere with MC4R activity, resulting in obesity in human patients.
Early SAR studies to improve selectivity and potency of melanocortin peptides focused mostly on MSH analogs (16, 17, 18, 37). In addition, Peng et al. (38) examined the role of C-terminal peptide residues of MSH and MSH in imparting selectivity of the peptides toward MCRs. They concluded that proline12 (numbering with respect to MSH) played an important role for binding and activity at the MC1R but not at MC3R. More recently Kavarana et al. (39) discovered that another cyclic MSH peptide analog, (O)C-CH2-CH2-c[His6-D-Phe7-Arg8-Trp9-Lys10]amide, was a selective and potent agonist for human MC4R. Furthermore, Balse-Schnivasan et al. (40) synthesized novel MSH/MSH hybrid analogs that led to potent and selective ligands to human MC3R and MC4R. Finally, Oosterom et al. (41, 42) reported that a major determinant for selective receptor interaction is the conformational presentation of the core sequence in melanocortin peptides directed to the receptor binding pockets of the MC3R and MC4R.
Holder and Haskell-Luevano (43) recently published a comprehensive review covering 30 yr of SAR studies on melanocortin ligands. However, almost all previous SAR studies were focused on MSH analogs. Because we believe that MSH(5–22) is an important endogenous ligand for MC4R and MC3R, we decided to perform our SAR based on the MSH(5–22) structure. From a relatively focused SAR study on MSH(5–22) described in this report, we discovered that the unique Arg6 in MSH (5–22) (or Arg2 in peptide 6 as an equivalent residue) plays a crucial role in maintaining the ideal conformation of the cyclic peptide so that it can interact potently and selectively with the binding sites of MC4R and MC3R. We believe that further development of the potent and selective cyclic MSH-derived peptides such as peptide 6 [Ac-YRcyclo(CEHdFRWC)amid]e may lead to more novel and superior MC4R peptide agonists, useful for treating obesity or metabolic syndrome.
One major focus of this investigation was to study the acute and chronic effects of the selective MC4R/MC3R agonist (peptide 6) on energy balance (food intake, fat use) and understand its mechanism of action. To achieve this goal, we administered peptide 6 either centrally (ICV) or peripherally (sc) to rats and found that peptide 6 reduced food intake and increased fat use in rats, irrespective of their route of administration or the fat contents in diets. In addition, sc administration of a high dose (30 μmol/kg) of peptide 6 to mice reduced food intake and increased fat use only in WT but not MC4R-deficient mice. The data suggest the peptide 6 reduced energy balance in rodents, mostly through its specific effects on MC4R that is expressed almost exclusively in brain and hypothalamus. These acute biological data were also consistent with the data obtained from the chronicle administration of the peptide dosed peripherally in male DIO rats (Figs. 1 and 2). In addition, peptide 6 was also found to cause the stretching behaviors (Emmerson, P., et al., personal communication) similar to what was observed for MTII, an MC4R peptide agonist. The combined data strongly suggest an important role of central MC4R mechanism of action for the biological activity of peptide 6.
We did not perform a conditioned taste aversion test with peptide 6. However, we do not believe peptide 6 induced aversion behavior because direct or indirect activation of MC4R with central administration of MTII, MSH, or leptin does not cause taste aversion in rats at the dose that caused anorectic effects (44). Furthermore, peptide 6 caused no changes in locomotor activity or total EEs in treated rats (Fig. 1, D and E), suggesting no toxic effects of the compound. Finally, peptide 6 reduced food intake and fat use specifically on WT mice but not MC4R knockout mice, suggesting the effects on body weight and food intake were not mediated through nonspecific aversive or toxic effects of the compound but through the specific effects on MC4R.
Peptide 6 has also been observed to have extended half-life (>>2 h) in rat and human sera and have greater than 80% bioavailability in rats and dogs when administered sc (data not shown). Our data suggest that the potent in vivo activity of peptide 6 when administered peripherally is due to not only its intrinsic high efficacy for MC4R but also its great stability and bioavailability of the peptide in the treated animals.
To conclude, we discovered, through a focused SAR study, a novel and selective MC4R and MC3R peptide agonist, peptide 6. This synthetic cyclic peptide administered either centrally or peripherally can potently reduce energy balance, increase fat use, and decrease weight gain in acute and chronic metabolic studies in rats. This antiobesity effect by peptide 6 was manifested only in WT but not MC4R-deficient mice, indicating that the antiobesity effect was attributed mostly through the MC4R agonist activity of the peptide.
Acknowledgments
The authors are indebted to Drs. Jose Caro and Paul Burn for encouragement in this study and Dr. Paul Emmerson, Dr. Bob Gadski, and Mr. Steve Kahl for helpful suggestions. The authors are grateful to the technical support of Mr. Dave Flora, Ms. Xiaoying Gao, and Mr. Tom P. O’Brien.
Footnotes
First Published Online September 15, 2005
Abbreviations: DIO, Diet-induced obese; EE, energy expenditure; ICV, intracerebroventricularly; Ki, affinity constant obtained from competitive binding assays; LPH, -lipotropin; MCR, melanocortin receptor; MTII, Ac-Nle-cyclo(DHdFRWK)amide; NDP, Nle4, D-Phe7; NMR, nuclear magnetic resonance; PC, prohormone convertase; POMC, proopiomelanocortin; RQ, respiratory quotient; SAR, structure-activity relationship; WT, wild type.
Accepted for publication September 9, 2005.
References
Pritchard LE, Turnbull AV, White A 2002 Pro-opiomelanocortin processing in the hypothalamus: impact on melanocortin signalling and obesity. J Endocrinol 172:411–421
Cone RD, Lu D, Koppula S, Vage DI, Klungland H, Boston B, Chen W, Orth DN, Pouton C, Kesterson RA 1996 The melanocortin receptors: agonists, antagonists, and the hormonal control of pigmentation. Rec Prog Horm Res 51:287–317; discussion 318
Bennett HP 1986 Biosynthetic fate of the amino-terminal fragment of pro-opiomelanocortin within the intermediate lobe of the mouse pituitary. Peptides 7:615–622
Adan RA, Gispen WH 2000 Melanocortins and the brain: from effects via receptors to drug targets. Eur J Pharmacol 405:13–24
O’Rahilly S, Farooqi IS, Yeo GS, Challis BG 2003 Minireview: human obesity-lessons from monogenic disorders. Endocrinology 144:3757–3764
Yeo GS, Lank EJ, Farooqi IS, Keogh J, Challis BG, O’Rahilly S 2003 Mutations in the human melanocortin-4 receptor gene associated with severe familial obesity disrupts receptor function through multiple molecular mechanisms. Hum Mol Genet 12:561–574
MacNeil DJ, Howard AD, Guan X, Fong TM, Nargund RP, Bednarek MA, Goulet MT, Weinberg DH, Strack AM, Marsh DJ, Chen HY, Shen CP, Chen AS, Rosenblum CI, MacNeil T, Tota M, MacIntyre ED, Van der Ploeg LH 2002 The role of melanocortins in body weight regulation: opportunities for the treatment of obesity. Eur J Pharmacol [Erratum (2002) 450:93–109] 440:141–157
Roselli-Rehfuss L, Mountjoy KG, Robbins LS, Mortrud MT, Low MJ, Tatro JB, Entwistle ML, Simerly RB, Cone RD 1993 Identification of a receptor for melanotropin and other proopiomelanocortin peptides in the hypothalamus and limbic system. Proc Natl Acad Sci USA 90:8856–8860
Bray GA 1993 The nutrient balance hypothesis: peptides, sympathetic activity, and food intake. Ann NY Acad Sci 676:223–241
Harrold JA, Widdowson PS, Williams G 2003 MSH: a functional ligand that regulated energy homeostasis via hypothalamic MC4-R Peptides 24:397–405
Harrold JA, Williams G, Widdowson PS 1999 Changes in hypothalamic agouti-related protein (AGRP), but not MSH or pro-opiomelanocortin concentrations in dietary-obese and food-restricted rats. Biochem Biophys Res Commun 258:574–577
Hsiung HM, Smiley DL, Zhang X-Y, Zhang L, Yan LZ, Craft L, Heiman ML, Smith DP 2005 Potent peptide agonists for human melanocortin 3/4-receptors derived from enzymatic cleavages of human MSH(5–22) by dipeptidyl peptidase I and dipeptidyl peptidase IV. Peptides 26:1988–1996
Challis BG, Pritchard LE, Creemers JW, Delplanque J, Keogh JM, Luan J, Wareham NJ, Yeo GS, Bhattacharyya S, Froguel P, White A, Farooqi IS, O’Rahilly S 2002 A missense mutation disrupting a dibasic prohormone processing site in pro-opiomelanocortin (POMC) increases susceptibility to early-onset obesity through a novel molecular mechanism. Hum Mol Genet 11:1997–2004
Bertagna X, Lenne F, Comar D, Massias JF, Wajcman H, Baudin V, Luton JP, Girard F 1986 Human -melanocyte-stimulating hormone revisited. Proc Natl Acad Sci USA 83:9719–9723
Krude H, Biebermann H, Gruters A 2003 Mutations in the human proopiomelanocortin gene. Ann NY Acad Sci 994:233–239
Sawyer TK, Hadley ME, Hruby VJ, Castrucci AMD, Staples DJ, Farah J, O’Donohue TL 1988 -Melanocyte-stimulating hormone structure-activity studies: comparative analysis of melanotropic and CNS bioactivities. Synapse 2:288–292
Sawyer TK, Hruby VJ, Darman PS, Hadley ME 1982 [half-Cys4,half-Cys10]--melanocyte-stimulating hormone: a cyclic -melanotropin exhibiting superagonist biological activity. Proc Natl Acad Sci USA 79:1751–1755
Hruby VJ, Lu D, Sharma SD, Castrucci AL, Kesterson RA, al-Obeidi FA, Hadley ME, Cone RD 1995 Cyclic lactam -melanotropin analogues of Ac-Nle4-cyclo[Asp5, D-Phe7,Lys10] -melanocyte-stimulating hormone-(4–10)-NH2 with bulky aromatic amino acids at position 7 show high antagonist potency and selectivity at specific melanocortin receptors. J Med Chem 38:3454–3461
Merrifield RB 1963 Solid phase peptide synthesis. I. The synthesis of a tetrapeptide. J Am Chem Soc 85:2149–2154
Wang SS 1973 p-Alkoxybenzyl alcohol resin and p-alkoxybenzyloxycarbonylhydrazide resin for solid phase synthesis of protected peptide fragments. J Am Chem Soc 95:1328–1333
Rink H 1987 Solid-phase synthesis of protected peptide fragments using a trialkoxy-diphenyl-methylester resin. Tetrahedron Lett 28:3787–3790
Berg DT, Grinnell BW, McClure DB 1993 High-level expression of secreted proteins from cells adapted to serum-free suspension culture. Biotechniques 14:972–978
Chen Y, Heiman ML 2000 Chronic leptin administration promotes lipid utilization until fat mass is greatly reduced and preserves lean mass of normal female rats. Regul Pept 92:113–119
Flatt J 1991 Assessment of daily and cumulative carbohydrate and fat balances in mice. J. Nutr Biochem 2:193–202
Chen Y, Hu C, Hsu CK, Zhang Q, Bi C, Asnicar M, Hsiung HM, Fox N, Slieker LJ, Yang DD, Heiman ML, Shi Y 2002 Targeted disruption of the melanin-concentrating hormone receptor-1 results in hyperphagia and resistance to diet-induced obesity. Endocrinology 143:2469–2477
Liu J, Na S, Glasebrook A, Fox N, Solenberg PJ, Zhang Q, Song HY, Yang DD 2001 Enhanced CD4+ T cell proliferation and Th2 cytokine production in DR6-deficient mice. Immunity 15:23–34
Tinsley FC, Taicher GZ, Heiman ML 2004 Evaluation of a quantitative magnetic resonance method for mouse whole body composition analysis. Obes Res 12:150–160
Sawyer TK, Sanfilippo PJ, Hruby VJ, Engel MH, Heward CB, Burnett JB, Hadley ME 1980 4-Norleucine, 7-D-phenylalanine--melanocyte-stimulating hormone: a highly potent -melanotropin with ultralong biological activity. Proc Natl Acad Sci USA 77:5754–5758
Cody WL, Mahoney M, Knittel JJ, Hruby VJ, Castrucci AM, Hadley ME 1985 Cyclic melanotropins. 9. 7-D-phenylalanine analogues of the active-site sequence. J Med Chem 28:583–588
Hirsch MD, O’Donohue TL, Wilson R, Sawyer TK, Hruby VJ, Hadley ME, Cody WL, Knittel JJ, Crawley JN 1984 Structural and conformational modifications of MSH/ACTH4–10 provide melanotropin analogues with highly potent behavioral activities. Peptides 5:1197–1201
Ying J, Kover KE, Gu X, Han G, Trivedi DB, Kavarana MJ, Hruby VJ 2003 Solution structures of cyclic melanocortin agonists and antagonists by NMR. Biopolymers 71:696–716
Holder JR, Haskell-Luevano C 2003 Melanocortin tetrapeptides modified at the N-terminus, His, Phe, Arg, and Trp positions. Ann NY Acad Sci 994:36–48
Holder JR, Bauzo RM, Xiang Z, Scott J, Haskell-Luevano C 2003 Design and pharmacology of peptoids and peptide-peptoid hybrids based on the melanocortin agonists core tetrapeptide sequence. Bioorg Med Chem Lett 13:4505–4509
Bertagna X 1994 Proopiomelanocortin-derived peptides. Endocrinol Metab Clin North Am 23:467–485
Raffin-Sanson ML, de Keyzer Y, Bertagna X 2003 Proopiomelanocortin, a polypeptide precursor with multiple functions: from physiology to pathological conditions. Eur J Endocrinol 149:79–90
Krude H, Biebermann H, Schnabel D, Tansek MZ, Theunissen P, Mullis PE, Gruters A 2003 Obesity due to proopiomelanocortin deficiency: three new cases and treatment trials with thyroid hormone and ACTH4–10. J Clin Endocrinol Metab 88:4633–4640
Danho W, Swistok J, Cheung AW, Chu XJ, Wang Y, Chen L, Bartkovitz D, Gore V, Qi L, Fry D, Greeley D, Sun H, Guenot J, Franco L, Kurylko G, Rumennik L, Yagaloff K 2001 Highly selective cyclic peptides for human melanocortin-4 receptor (MC4-R): design, synthesis, bioactive conformation, and pharmacological evaluation as anti-obesity agent. Peptides: the wave of the future. Proc Second International and the Seventeenth American Peptide Symposium, San Diego, CA, 2001, pp 701–703
Peng PJ, Sahm UG, Doherty RV, Kinsman RG, Moss SH, Pouton CW 1997 Binding and biological activity of C-terminally modified melanocortin peptides: a comparison between their actions at rodent MC1 and MC3 receptors. Peptides 18:1001–1008
Kavarana MJ, Trivedi D, Cai M, Ying J, Hammer M, Cabello C, Grieco P, Han G, Hruby VJ 2002 Novel cyclic templates of MSH give highly selective and potent antagonists/agonists for human melanocortin-3/4 receptors. J Med Chem 45:2644–2650
Balse-Srinivasan P, Grieco P, Cai M, Trivedi D, Hruby VJ 2003 Structure-activity relationships of MSH analogues at the human melanocortin MC3, MC4, and MC5 receptors. Discovery of highly selective hMC3R, hMC4R, and hMC5R analogues. J Med Chem 46:4965–4973
Oosterom J, Nijenhuis WA, Schaaper WM, Slootstra J, Meloen RH, Gispen WH, Burbach JP, Adan RA 1999 Conformation of the core sequence in melanocortin peptides directs selectivity for the melanocortin MC3 and MC4 receptors. J Biol Chem 274:16853–16860
Oosterom J, Garner KM, den Dekker WK, Nijenhuis WA, Gispen WH, Burbach JP, Barsh GS, Adan RA 2001 Common requirements for melanocortin-4 receptor selectivity of structurally unrelated melanocortin agonist and endogenous antagonist, agouti protein. J Biol Chem 276:931–936
Holder JR, Haskell-Luevano C 2004 Melanocortin ligands: 30 years of structure-activity relationship (SAR) studies. Med Res Rev 24:325–356
Wirth MM, Olszewski PK, Yu C, Levine AS, Giraudo SQ 2001 Paraventricular hypothalamic -melanocyte-stimulating hormone and MTII reduce feeding without causing aversive effects. Peptides 22:129–134(Hansen M. Hsiung, JeAnne Hertel, Xing-yu)