Functional Mixture-Based Positional Scan Identifies a Library of Antagonist Tetrapeptide Sequences (LAtTeS) with Nanomolar Potency for the Melanocortin-4 Receptor and Equipotent with the Endogenous AGRP(86–132) Antagonist

Abstract

The melanocortin-4 receptor (MC4R) plays an important role in appetite. Agonist ligands that stimulate the MC4R decrease appetite, while antagonist compounds increase food consumption. Herein, a functional mixture-based positional scan identified novel MC4R antagonist sequences. Mixtures comprising a library of 12,960,000 tetrapeptides were screened in the presence and absence of NDP-MSH agonist. These results led to the synthesis of forty-eight individual tetrapeptides, of which forty were screened for functional activity at the melanocortin receptors. Thirteen compounds were found to possess nanomolar antagonist potency at the MC4R, with the general tetrapeptide sequence Ac-Aromatic-Basic-Aromatic-Basic-NH₂. The most notable results include the identification of tetrapeptide 48 [COR1–25, Ac-DPhe(pI)-Arg-Nal(2’)-Arg-NH₂], an equipotent MC4R antagonist to agouti-related protein [AGRP(86–132)], more potent than miniAGRP(87–120), and possesses 15-fold selectivity for the MC4R versus the MC3R. These tetrapeptides may serve as leads for novel appetite-inducing therapies to treat states of negative energy balance, such as cachexia and anorexia.

Graphical Abstract

Introduction

The melanocortin system is comprised of five G protein-coupled receptors,^1–8 endogenous agonists (including α-MSH, β-MSH, γ-MSH, and ACTH) derived from the proopiomelanocortin (POMC) gene transcript,⁹ and the naturally occurring antagonists agouti signaling protein (ASP)^10–11 and agouti-related protein (AGRP).^12–14 The melanocortin receptor (MCR) family is involved in many physiological processes, including pigmentation (MC1R),^{2, 7} steroidogenesis (MC2R),^{7, 15} exocrine gland function in rodents (MC5R),^16–17 and energy homeostasis/food intake/obesity/anorexia (MC3R and MC4R).^18–21 Selective knockout of the MC4R in mice results in hyperphagia and obesity,¹⁹ with a similar phenotype observed in humans deficient in MC4R signaling.^22–24 As the MC4R is one of the central regulators of appetite and body weight, selective ligands for this receptor are highly desirable. Purported MC4R-selective agonist peptide ligands (LY2112688²⁵ and setmelanotide^26–29) have been reported to decrease appetite and food consumption in human trials, although LY2112088 was not advanced due to cardiovascular liabilities.²⁵ Trials in limited patient populations resulted in FDA approval of setmelanotide (Imcivree) in 2020 for chronic weight management for individuals with genetically-confirmed pathogenic or uncertain significance variants in POMC, proprotein convertase subtilisin/kexin type 1 (PCSK1), or leptin receptor (LEPR) genes.

While central administration of melanocortin agonist ligands decreases food intake, dosing of MC4R antagonist ligands has been demonstrated to increase food intake in rodents.^{18, 20, 30–31} Compounds that increase appetite may be useful to treat diseases of negative energy balance such as anorexia and cachexia associated with cancer.^32–35 Pioneering work by Hruby et al. demonstrated that substituting the amino acid DNal(2’) for DPhe converted the synthetic and potent melanocortin agonists NDP-MSH and MTII into MC3R/MC4R antagonist peptides.^36–37 These data indicate that structural modifications to known melanocortin agonist compounds can generate antagonist ligands. Another approach to generate novel MC4R antagonist ligands may be to modify the naturally occurring MC4R antagonist ligands ASP and AGRP. Although the large size of active forms of ASP [ASP(23–131)]³⁸ and AGRP [AGRP(83–132)]³⁹ limit their feasibility as therapeutic leads, it was previously reported that the hypothesized hexapeptide active loop of AGRP could be cyclized through a DPro-Pro motif.⁴⁰ This resulted in the macrocyclic octapeptide c[Pro-Arg-Phe-Phe-Asn-Ala-Phe-DPro] that possessed sub-micromolar functional potency at the MC4R.⁴⁰ Further studies identified substitutions that resulted in sub-nanomolar potency at the MC4R with 600- to 800-fold selectivity for the MC4R over the MC3R.⁴¹ As this scaffold is similar in size to the cyclic, FDA-approved setmelanotide agonist compound, modifications to the naturally occurring melanocortin antagonists represent another strategy for discovering MC4R antagonist ligands.

Although the rational design of MC4R antagonist ligands has generated potent and selective compounds, the resulting antagonists are limited by known scaffolds and structure-activity relationships, which may introduce assumptions in the design process. An unbiased method exploring a large area of chemical space may identify new scaffolds, leading to ligands that are not readily apparent from previously known compounds or activity trends. Mixture-based positional scanning libraries are one technique that can identify novel active compounds, as reviewed.^42–44 Rescue compounds for MC4R polymorphisms that are properly expressed at the cell surface but do not respond to the endogenous melanocortin ligands,⁴⁵ agonist peptides selective for MC3R over MC4R,^46–47 and small molecule MC3R ligands⁴⁸ have all been identified using this technology. These melanocortin examples all involved identifying new agonist compounds. Since the melanocortin receptors couple to G_αS protein subunits and increase cAMP levels upon receptor activation,⁴⁹ the discovery of agonist compounds utilizes in vitro gain-of-signal cellular assays to monitor cAMP increases in our laboratory. As antagonists do not possess functional activity in isolation, antagonist functional mixture-based positional scanning requires the addition of an agonist ligand to create a signal, which is decreased in the presence of an antagonist mixture in gain-of-signal assays.

This more complex antagonist mixture-based positional scanning approach has been performed in mouse B16-F10 melanoma cells (presumably expressing the MC1R) to identify ligands that antagonize α-MSH-stimulated melanin production.⁵⁰ A tetrapeptide library composed of the natural twenty amino acids at each position (160,000 theoretical compounds in eighty mixtures) was screened in B16-F10 cells for anti-melanogenic activity. Cells were pre-treated with a peptide mixture concentration of 1 mM, followed by stimulation with 100 nM α-MSH.⁵⁰ After 72 h, the melanin concentration was measured. Following mixture deconvolution and individual compound synthesis, two tetrapeptides (H-RFWG-NH₂ and H-RLWG-NH₂) were identified that inhibited α-MSH-stimulated melanin production at concentrations of 30 and 100 μM.⁵⁰ Further truncation studies led to the observation that the C-terminal amino acid in these tetrapeptides, glycinamide (H-G-NH₂), possessed anti-melanogenic activity at 30 and 100 μM concentrations.⁵⁰ Follow-up work in a human clinical trial reported that application of glycinamide to facial skin decreased melanin index and skin pigmentation, while increasing skin lightness, as compared to a control formulation lacking glycinamide.⁵¹ These data suggest that an in-vitro functional antagonist mixture-based positional scan can successfully identify clinically relevant compounds.

To identify novel MC4R peptide antagonists that might serve as lead compounds in the development of weight-gain therapeutics, a mixture-based positional scan was performed using a tetrapeptide library at the mouse (m)MC4R. The mouse versus the human MC4R was selected so that novel peptides discovered herein could be used in the future to directly correlate with in vivo mouse experimental physiological data without the need for additional in vitro experiments. Sixty individual building blocks were incorporated at each position, resulting in 240 mixtures (containing 12,960,000 tetrapeptides) assayed at the mMC4R using the synthetic peptide NDP-MSH as the agonist (Figure 1). Following the initial screen at 50 μg/mL concentrations, the mixtures that resulted in the highest putative antagonist potency were re-screened at concentrations of 25 and 50 μg/mL. Based upon these data, substitutions for each of the four tetrapeptide positions were selected, resulting in the deconvolution and synthesis of a forty-eight-member library. Forty tetrapeptides were soluble in solvents compatible with RP-HPLC and were subsequently assayed for agonist activity at the mMC1R, mMC3R, mMC4R, and mMC5R, as well as for antagonist activity at the mMC3R and mMC4R. The remaining eight tetrapeptides, compromised of aromatic amino acids at each position, were insoluble as anticipated, i.e. Ac-DNal(2’)-DNal(2’)-DNal(2’)-DNal(2’)-NH₂.

Figure 1:

Experimental workflow to describe the current results. Both an agonist and antagonist mixture-based positional scan was performed at the mMC4R, utilizing a 12,960,000 tetrapeptide library contained in 240 mixtures. Twenty-six mixtures with putative antagonist potency, representing 5,616,000 tetrapeptides, were re-assayed using two different concentrations as agonists and antagonists. Following deconvolution of these results, forty-eight individual peptides were synthesized and the soluble forty tetrapeptides were assayed at the mMC1R, mMC3R, mMC4R, and mMC5R. Of these, thirteen tetrapeptides possessed nanomolar potent MC4R antagonist activity (pA₂ > 8).

A common concern to using this “unbiased” strategy is that the “most active” compound from a mixture-based positional scanning library may not be identified. This is a valid concern to which one may never truly know the answer. Another caveat of utilizing any library screening approach is that it is limiting based upon the building blocks incorporated, in this case the amino acids incorporated into the various positions of the tetrapeptide template. The present library screened a theoretical 12,960,000 tetrapeptides and yielded a particular scaffold (Ac-Aromatic-Basic-Aromatic-Basic-NH₂) that resulted in nanomolar potent MC4R antagonists. As with any study, future structure activity relationship experiments incorporating other amino acids not present in the current mixtures may yield the discovery of more potent ligands.

Results & Discussion

Primary Mixture-Based Positional Screen

The mixture-based tetrapeptide positional scanning library TPI924, possessing an N-terminal acetyl group and C-terminal amide, was constructed using standard Boc chemistry and the previously reported tea-bag method for generating compound mixtures.⁵² Each mixture within the library held one position constant while varying the other three positions with sixty natural and unnatural amino acids, resulting in 216,000 (1 × 60 × 60 × 60) tetrapeptides per mixture. The overall library consisted of 12,960,000 (60 × 60 × 60 × 60) tetrapeptides in 240 mixtures, which has previously been used to identify MC3R and MC4R agonist ligands.^45–47

As an initial screen, the 240 mixtures were assayed in both agonist and antagonist paradigms using a modified β-galactosidase assay⁵³ with the fluorescent substrate 4-methylumbelliferyl-β-D-galactopyranoside to measure cAMP production in HEK293 cells stably expressing the mMC4R. The agonist screen utilized a mixture concentration of 50 μg/mL to identify which mixtures could induce an increase in cAMP in cells expressing the mMC4R (Table S1). The established positive control tetrapeptide Ac-His-DPhe-Arg-Trp-NH₂ possesses nanomolar agonist potency at the mMC4R.⁵⁴ Mixtures possessing this tetrapeptide sequence (His at position R1, DPhe at position R2, Arg at position R3, and Trp at position R4) resulted in a significant agonist response (68–81% compared to the maximal NDP-MSH signal, Table S1) for each of these four mixtures, as anticipated, indicating the screen possessed the requisite sensitivity to identify mMC4R agonist sequences.

The 240 mixtures were also screened in an antagonist paradigm at a single concentration (50 μg/mL) in the presence of 0.5 nM NDP-MSH (Figure 2 & Table S1). Each well fluorescent signal was normalized to well protein level (cell number control), plate basal signal (negative control), and plate maximal NDP-MSH signal (positive control). The normalized signal was converted to a percentage from the plate median experimental value to basal signal (Figure 2 & Table S1) as an approximation for putative mixture antagonist potency. For each position, the mixtures are arranged in order of observed putative antagonist activity at 50 μg/mL concentrations (Figure 2). The five to eight mixtures that possessed the highest putative antagonist potency at each position (black dots in Figure 2, marked as selected for follow-up in Table S1) were chosen for rescreening using two concentrations to confirm putative antagonist activity and identify the most active mixtures.

Figure 2:

Illustration of the primary antagonist screening results of the TPI924 library at the mMC4R. The TPI924 library (240 mixtures) was screened at 50 μg/mL concentrations in the presence of 0.5 nM NDP-MSH in a cAMP based fluorescent β-galactosidase assay. The fluorescence signal (indicative of cAMP levels), was normalized to well protein levels, plate basal signal, and plate maximal NDP-MSH signal. The normalized signal was converted to a percentage from the median plate experimental value to basal signal. A larger percentage corresponds to more potent putative antagonist activity. The X-axes represent the amino acid residue that was held constant at that position (O) of a tetrapeptide library with the three remaining positions composed of 60 amino acids. Mixtures are arranged by putative antagonist potency. Black dots indicate mixtures that were selected for a follow-up two-point concentration-response assay.

The follow-up screen was performed using twenty-six compound mixtures at two concentrations (25 and 50 μg/mL) in both an agonist (mixture by itself) and antagonist (mixture with 0.5 nM NDP-MSH) paradigm (Figure 3 and Table S2). In the agonist assays, all twenty-six mixtures resulted in less than 40% activity compared to the maximal response of NDP-MSH at the highest 50 μg/mL concentration assayed (Table S2). For the antagonist screen, following normalization of the fluorescent signal to protein, basal, and maximal NDP-MSH, the signal for each mixture was converted to a percentage from a control (no mixture) 0.5 nM NDP-MSH signal to basal activity (Figure 3). Within this approach, larger values indicate more potent putative antagonist activity. Both the 50 (blue bars) and 25 (red bars) μg/mL mixture concentrations in the presence of 0.5 nM NDP-MSH are presented in Figure 3, with the mixtures group by position and arranged in observed putative antagonist potency at the 25 μg/mL concentration. Select mixtures did not result in putative antagonism in the follow-up screen (DNle and DTic at the R2 position, DPhe and DGlu at the R4 position). Of the remaining twenty-two mixtures, seventeen [DNal(2’), DPhe(pI), Cha, Phe(pI), and Trp at R1, DNal(2’), Arg, and Ile at R2, DNal(2’), Trp, Nal(2’), Nle, Phe(pI), and Tyr at R3, and DLys, DNle, and DTic at R4] resulted in a concentration-dependent decrease in the putative antagonist signal [the observed antagonist signal was stronger at the 50 μg/mL concentration (blue bars) as compared to the 25 μg/mL concentration (red bars)], indicating that the putative antagonist signals were dose dependent. The defined amino acids of the mixtures showing the most potent putative antagonist activity at the 25 μg/mL concentration for each position were selected for individual compound synthesis, indicated by black dots on Figure 3 and summarized in Table 1 and Figure 4. Individual compounds may be classified into one of four general amino acid substitution patterns: Ac-Aromatic-Aromatic-Aromatic-Aromatic-NH₂, Ac-Aromatic-Basic-Aromatic-Aromatic-NH₂, Ac-Aromatic-Aromatic-Aromatic-Basic-NH₂, and Ac-Aromatic-Basic-Aromatic-Basic-NH₂.

Figure 3:

Illustration of the follow-up two-point antagonist concentration-response screen of the TPI924 library at the mMC4R. The antagonist screen was assayed using 0.5 nM NDP-MSH in the presence of 25 or 50 μg/mL TPI924 concentrations (red and blue bars, respectively). The fluorescence signal was normalized to well protein levels, plate basal signal, and plate maximal NDP-MSH signal. The normalized signal was converted to a percentage from the signal of control 0.5 nM NDP-MSH (without mixture) to basal signal. A higher percentage corresponds to more potent putative antagonist activity. The X-axis represents the amino acid residue that was held constant at that position (position indicated below) with the three remaining positions composed of 60 amino acids. Mixtures are grouped by position within the tetrapeptide scaffold and arranged by putative antagonist potency using the 25 μg/mL TPI924 with 0.5 nM NDP-MSH results (red bars). Black dots indicate amino acids at specific positions that were used in the library design of individual tetrapeptides.

Table 1:

Summary of the Proposed Tetrapeptide Residues Following Mixture-Based Positional Scanning Deconvolution

R1	R2	R3	R4
DNal(2’)	DNal(2’)	DNal(2’)	DLys
DPhe(pI)	Arg	Trp	DNal(2’)
		DPhe(pI)	Arg
		Nal(2’)

Figure 4:

Structures of the tetrapeptide amino acids used in this study.

Individual Compound Library

Individual tetrapeptides (Table S3) were synthesized manually using standard Fmoc chemical techniques.^55–56 All tetrapeptides were acetylated at the N-terminal and possessed a C-terminal carboxamide functionality. Following cleavage and sidechain deprotection, peptides were purified by semi-preparative RP-HPLC to greater than 95%. Eight tetrapeptides that were composed of four aromatic amino acids at each position (Ac-Aromatic-Aromatic-Aromatic-Aromatic-NH₂) were not soluble in solvents compatible with RP-HPLC and were not further investigated (Table S3, denoted as insoluble). For the remaining forty tetrapeptides, purity was assessed using analytical RP-HPLC (traces in the SI) and the molecular mass was determined by ESI-MS (University of Minnesota Mass Spectrometry Laboratory). The tetrapeptides were assessed for biological activity using the AlphaScreen cAMP assay at the mMC1R, mMC3R, mMC4R, and mMC5R stably expressed in HEK293 cells, as previously described.^57–59 Since the MC2R is only stimulated by ACTH, it was excluded from this study. Tetrapeptides were assayed for antagonist activity at the mMC4R using NDP-MSH as the agonist, as well as for agonist activity at the mMC1R, mMC3R, and mMC5R. Since none of the compounds were full agonists at the mMC3R at concentrations up to 100 μM, the forty soluble tetrapeptides were also assayed for antagonist activity at the mMC3R with the agonist NDP-MSH. Tetrapeptides that did not possess agonist or antagonist activity in two independent experimental replicates were not further characterized. Compounds were assayed in duplicate wells in at least three independent experiments if agonist or antagonist activity was observed. Tetrapeptides within a 3-fold potency range were considered equipotent due to the inherent error of the assay in our hands. Compounds that activated the receptor to 90% of the maximal signal of NDP-MSH were considered full agonists, while compounds that activated the receptor to less than 20% were considered inactive. Since the cAMP AlphaScreen assay is a loss-of-signal competition assay, in which higher concentrations of compound result in lower assay signal, the data is normalized to baseline and maximal NDP-MSH signal for illustrative purposes, as previously described.^60–61

Melanocortin-4 Receptor Activity:

The mixtures used to identify the compounds synthesized were screened for both agonist and antagonist activity and the deconvolution was carried out for antagonistic activity. Of the forty tetrapeptides that were purified and screened, none were able to fully stimulate the mMC4R at the highest concentration assayed (100 μM, Table 2), an expected result for minimal mMC4R agonist activity based upon the initial mixture screening and the selected amino acids for deconvolution. Thirteen of the forty tetrapeptides were able to partially activate the mMC4R at the highest concentration assayed (100 μM), with signals ranging from 20% to 60% of the maximal signal observed for NDP-MSH. An example of the partial activation is provided in Figure 5 for 48 (COR1-25), which stimulated the mMC4R to 25% of the maximal NDP-MSH signal at 100 μM concentrations.

Table 2:

Tetrapeptide Agonist Pharmacology at the Mouse Melanocortin Receptors.^a

Peptide	Compound ID	Sequence	mMC1R	mMC3R	mMC4R	mMC5R
			EC50 (nM)	EC50 (nM)	EC50 (nM)	EC50 (nM)
	NDP-MSH		0.012±0.001	0.074±0.006	0.47±0.03	0.21±0.01
AGRP(86–132) b			>100μM			>100μM
miniAGRP(87–120) b			>100μM			>100μM
1	COR1-29-1	Ac-DNal(2’)-DNal(2’)-DNal(2’)-DLys-NH2	55% @100μM	>100,000	>100,000	>100,000
3	COR2-99	Ac-DNal(2’)-DNal(2’)-DNal(2’)-Arg-NH2	30% @100μM	35% @100μM	25% @100μM	130±60
4	COR1-29-2	Ac-DNal(2’)-DNal(2’)-Trp-DLys-NH2	25% @100μM	>100,000	30% @100μM	>100,000
6	MDE10-29	Ac-DNal(2’)-DNal(2’)-Trp-Arg-NH2	45% @100μM	35% @100μM	60% @100μM	45% @100μM
7	COR1-29-3	Ac-DNal(2’)-DNal(2’)-DPhe(pI)-DLys-NH2	20% @100μM	>100,000	>100,000	>100,000
9	COR2-87	Ac-DNal(2’)-DNal(2’)-DPhe(pI)-Arg-NH2	25% @100μM	>100,000	>100,000	30±10
10	COR1-29-4	Ac-DNal(2’)-DNal(2’)-Nal(2’)-DLys-NH2	45% @100μM	>100,000	>100,000	>100,000
12	MDE10-63	Ac-DNal(2’)-DNal(2’)-Nal(2’)-Arg-NH2	55% @100μM	>100,000	>100,000	40% @100μM
13	COR1-29-5	Ac-DNal(2’)-Arg-DNal(2’)-DLys-NH2	Partial Agonist 6,000±4,000 (30% NDP)	>100,000	>100,000	>100,000
14	COR2-57	Ac-DNal(2’)-Arg-DNal(2’)-DNal(2’)-NH2	Partial Agonist 5,000±1,000 (85% NDP)	50% @100μM	45% @100μM	20% @100μM
15	COR2-93	Ac-DNal(2’)-Arg-DNal(2’)-Arg-NH2	Partial Agonist 1,000±500 (60% NDP)	>100,000	>100,000	>100,000
16	COR1-29-6	Ac-DNal(2’)-Arg-Trp-DLys-NH2	>100,000	>100,000	>100,000	>100,000
17	COR1-60	Ac-DNal(2’)-Arg-Trp-DNal(2’)-NH2	40% @100μM	40% @100μM	30% @100μM	>100,000
18	MDE10-27	Ac-DNal(2’)-Arg-Trp-Arg-NH2	40% @100μM	>100,000	20% @100μM	>100,000
19	COR1-29-7	Ac-DNal(2’)-Arg-DPhe(pI)-DLys-NH2	220±50	>100,000	>100,000	>100,000
20	COR2-15	Ac-DNal(2’)-Arg-DPhe(pI)-DNal(2’)-NH2	5,000±3,000	40% @100μM	30% @100μM	40% @100μM
21	COR2-105	Ac-DNal(2’)-Arg-DPhe(pI)-Arg-NH2	40±20	>100,000	>100,000	>100,000
22	COR1-29-8	Ac-DNal(2’)-Arg-Nal(2’)-DLys-NH2	Partial Agonist 3,300±900 (50% NDP)	>100,000	>100,000	>100,000
23	COR2-33	Ac-DNal(2’)-Arg-Nal(2’)-DNal(2’)-NH2	40% @100μM	30% @100μM	30% @100μM	25% @100μM
24	MDE10-28	Ac-DNal(2’)-Arg-Nal(2’)-Arg-NH2	>100,000	>100,000	30% @100μM	>100,000
25	COR1-80	Ac-DPhe(pI)-DNal(2’)-DNal(2’)-DLys-NH2	35% @100μM	>100,000	>100,000	>100,000
27	COR1-65	Ac-DPhe(pI)-DNal(2’)-DNal(2’)-Arg-NH2	>100,000	>100,000	>100,000	230±60
28	COR1-95	Ac-DPhe(pI)-DNal(2’)-Trp-DLys-NH2	65% @100μM	>100,000	>100,000	>100,000
30	COR1-90	Ac-DPhe(pI)-DNal(2’)-Trp-Arg-NH2	>100,000	>100,000	>100,000	>100,000
31	COR2-9	Ac-DPhe(pI)-DNal(2’)-DPhe(pI)-DLys-NH2	>100,000	>100,000	>100,000	>100,000
33	COR1-55	Ac-DPhe(pI)-DNal(2’)-DPhe(pI)-Arg-NH2	40% @100μM	>100,000	>100,000	130±50
34	COR2-117	Ac-DPhe(pI)-DNal(2’)-Nal(2’)-DLys-NH2	40% @100μM	20% @100μM	>100,000	50% @100μM
36	COR2-3	Ac-DPhe(pI)-DNal(2’)-Nal(2’)-Arg-NH2	20% @100μM	>100,000	>100,000	800±400
37	COR2-111	Ac-DPhe(pI)-Arg-DNal(2’)-DLys-NH2	Partial Agonist 3,500±900 (30% NDP)	>100,000	>100,000	>100,000
38	COR2-27	Ac-DPhe(pI)-Arg-DNal(2’)-DNal(2’)-NH2	80% @100μM	45% @100μM	45% @100μM	20% @100μM
39	COR1-19	Ac-DPhe(pI)-Arg-DNal(2’)-Arg-NH2	Partial Agonist 400±300 (65% NDP)	>100,000	>100,000	>100,000
40	COR1-50	Ac-DPhe(pI)-Arg-Trp-DLys-NH2	>100,000	>100,000	>100,000	>100,000
41	COR1-85	Ac-DPhe(pI)-Arg-Trp-DNal(2’)-NH2	45% @100μM	25% @100μM	35% @100μM	>100,000
42	COR1-7	Ac-DPhe(pI)-Arg-Trp-Arg-NH2	Partial Agonist 7900±500 (50% NDP)	25% @100μM	>100,000	>100,000
43	COR1-100	Ac-DPhe(pI)-Arg-DPhe(pI)-DLys-NH2	35±3	>100,000	>100,000	>100,000
44	COR1-105	Ac-DPhe(pI)-Arg-DPhe(pI)-DNal(2’)-NH2	1,100±500	40% @100μM	40% @100μM	65% @100μM
45	COR1-13	Ac-DPhe(pI)-Arg-DPhe(pI)-Arg-NH2	7±1	>100,000	>100,000	>100,000
46	COR1-32	Ac-DPhe(pI)-Arg-Nal(2’)-DLys-NH2	>100,000	>100,000	>100,000	>100,000
47	COR1-46	Ac-DPhe(pI)-Arg-Nal(2’)-DNal(2’)-NH2	55% @100μM	>100,000	>100,000	>100,000
48	COR1-25	Ac-DPhe(pI)-Arg-Nal(2’)-Arg-NH2	25% @100μM	>100,000	25% @100μM	>100,000
49	RHC1-65	Ac-Phe(pI)-Nal(2’)-Phe(pI)-Arg-NH2	60% @100μM	>100,000	>100,000	>100,000

^aThe indicated errors represent the standard error of the mean determined from at least three independent experiments. A percentage denotes the percent maximal stimulatory response (compared to NDP-MSH) observed at 100 μM, but not enough stimulation was observed to determine an EC₅₀ value. The use of >100,000 indicates that the compound was examined but lacked agonist activity at concentrations up to 100 μM in at least two independent experiments. Tetrapeptides labeled partial agonists were observed to possess partial agonist activity, with the apparent EC₅₀ values and percentage of receptor activation relative to NDP-MSH. NA indicates that the tetrapeptide was not assayed as an antagonist.

^bThe AGRP(86–132) data used the same cAMP-based AlphaScreen assay herein.⁶² The miniAGRP(87–120) data are from our laboratory using a cAMP-based β-galactosidase reporter gene assay.^{40, 63}

Figure 5:

Illustration of the antagonist pharmacology of 48 (COR1-25) and 9 (COR2-87) at the mMC3R and mMC4R.

While minimal agonist activity was observed at the mMC4R, thirteen compounds were nanomolar potent antagonists (pA₂ values between 8–9) at the mMC4R (Table 3). All thirteen tetrapeptides possessed the general substitution pattern Ac-Aromatic-Basic-Aromatic-Basic-NH₂. The three remaining tetrapeptides with this substitution pattern, 16 (COR1-29-6), 22 (COR1-29-8), and 40 (COR1-50) possessed pA₂ values of 7.7, 7.8, and 7.6 respectively. Only one compound outside of this substitution pattern that had a pA₂ value greater than 7 [41 (COR1-85), pA₂ = 7.6]. Overall the Ac-Aromatic-Basic-Aromatic-Basic-NH₂ scaffold yielded the fifteen most potent mMC4R antagonist tetrapeptides in this study. As this scaffold possesses two basic residues compared to one basic residue in either the Ac-Aromatic-Basic-Aromatic-Aromatic-NH₂ or Ac-Aromatic-Aromatic-Aromatic-Basic-NH₂ substitution patterns, the more potent mMC4R antagonists may also be more hydrophilic. Using the HPLC k’ (ACN) value as an approximation for hydrophilicity (smaller value = more hydrophilic), plotting the HPLC k’ (ACN) versus mMC4R pA₂ values results in two distinct compound groups (Figure S1). The Ac-Aromatic-Basic-Aromatic-Basic-NH₂ pattern groups as a more potent, more hydrophilic cluster, while the single basic substitution patterns possess longer retention times (more hydrophobic) and lower mMC4R antagonist potencies.

Table 3:

Tetrapeptide Antagonist Pharmacology at the Mouse Melanocortin Receptors.^a

Peptide	Compound ID	Sequence	mMC3R		mMC4R
			pA2	Antagonist Potency Ki (nM)	pA2	Antagonist Potency Ki (nM)
AGRP(86–132) b			8.7±0.1	2.0	8.7±0.2	2.0
miniAGRP(87–120) b			8.1±0.1	7.9	8.5±0.1	3.2
1	COR1-29-1	Ac-DNal(2’)-DNal(2’)-DNal(2’)-DLys-NH2	<5.5		<5.5
3	COR2-99	Ac-DNal(2’)-DNal(2’)-DNal(2’)-Arg-NH2	<5.5		<5.5
4	COR1-29-2	Ac-DNal(2’)-DNal(2’)-Trp-DLys-NH2	5.8±0.1	1600	6.7±0.3	200
6	MDE10–29	Ac-DNal(2’)-DNal(2’)-Trp-Arg-NH2	<5.5		<5.5
7	COR1-29-3	Ac-DNal(2’)-DNal(2’)-DPhe(pI)-DLys-NH2	<5.5		<5.5
9	COR2-87	Ac-DNal(2’)-DNal(2’)-DPhe(pI)-Arg-NH2	<5.5		<5.5
10	COR1-29-4	Ac-DNal(2’)-DNal(2’)-Nal(2’)-DLys-NH2	<5.5		<5.5
12	MDE10–63	Ac-DNal(2’)-DNal(2’)-Nal(2’)-Arg-NH2	<5.5		<5.5
13	COR1-29-5	Ac-DNal(2’)-Arg-DNal(2’)-DLys-NH2	6.0±0.1	1000	8.2±0.2	6.3
14	COR2-57	Ac-DNal(2’)-Arg-DNal(2’)-DNal(2’)-NH2	5.7±0.1	2000	6.3±0.1	500
15	COR2-93	Ac-DNal(2’)-Arg-DNal(2’)-Arg-NH2	6.1±0.1	790	8.3±0.3	5.0
16	COR1-29-6	Ac-DNal(2’)-Arg-Trp-DLys-NH2	5.5±0.1	3200	7.7±0.3	20
17	COR1-60	Ac-DNal(2’)-Arg-Trp-DNal(2’)-NH2	5.7±0.1	2000	7.0±0.1	100
18	MDE10–27	Ac-DNal(2’)-Arg-Trp-Arg-NH2	6.8±0.3	160	8.3±0.1	5.0
19	COR1-29-7	Ac-DNal(2’)-Arg-DPhe(pI)-DLys-NH2	6.9±0.1	130	8.5±0.3	3.2
20	COR2-15	Ac-DNal(2’)-Arg-DPhe(pI)-DNal(2’)-NH2	6.1±0.1	790	6.5±0.1	320
21	COR2-105	Ac-DNal(2’)-Arg-DPhe(pI)-Arg-NH2	7.1±0.1	79	8.6±0.3	2.5
22	COR1-29-8	Ac-DNal(2’)-Arg-Nal(2’)-DLys-NH2	6.2±0.1	630	7.8±0.3	16
23	COR2-33	Ac-DNal(2’)-Arg-Nal(2’)-DNal(2’)-NH2	<5.5		6.7±0.1	200
24	MDE10–28	Ac-DNal(2’)-Arg-Nal(2’)-Arg-NH2	7.0±0.1	100	8.8±0.1	1.6
25	COR1-80	Ac-DPhe(pI)-DNal(2’)-DNal(2’)-DLys-NH2	<5.5		6.3±0.2	500
27	COR1-65	Ac-DPhe(pI)-DNal(2’)-DNal(2’)-Arg-NH2	<5.5		<5.5
28	COR1-95	Ac-DPhe(pI)-DNal(2’)-Trp-DLys-NH2	5.7±0.1	2000	6.6±0.1	250
30	COR1-90	Ac-DPhe(pI)-DNal(2’)-Trp-Arg-NH2	<5.5		5.6±0.1	2500
31	COR2-9	Ac-DPhe(pI)-DNal(2’)-DPhe(pI)-DLys-NH2	<5.5		<5.5
33	COR1-55	Ac-DPhe(pI)-DNal(2’)-DPhe(pI)-Arg-NH2	<5.5		<5.5
34	COR2-117	Ac-DPhe(pI)-DNal(2’)-Nal(2’)-DLys-NH2	<5.5		<5.5
36	COR2-3	Ac-DPhe(pI)-DNal(2’)-Nal(2’)-Arg-NH2	<5.5		5.8±0.1	1600
37	COR2-111	Ac-DPhe(pI)-Arg-DNal(2’)-DLys-NH2	7.2±0.2	63	8.6±0.1	2.5
38	COR2-27	Ac-DPhe(pI)-Arg-DNal(2’)-DNal(2’)-NH2	6.1±0.1	790	6.7±0.1	200
39	COR1-19	Ac-DPhe(pI)-Arg-DNal(2’)-Arg-NH2	6.9±0.2	130	8.3±0.1	5.0
40	COR1-50	Ac-DPhe(pI)-Arg-Trp-DLys-NH2	6.2±0.1	630	7.6±0.2	25
41	COR1-85	Ac-DPhe(pI)-Arg-Trp-DNal(2’)-NH2	6.0±0.1	1000	7.6±0.3	25
42	COR1-7	Ac-DPhe(pI)-Arg-Trp-Arg-NH2	7.4±0.1	40	8.7±0.1	2.0
43	COR1-100	Ac-DPhe(pI)-Arg-DPhe(pI)-DLys-NH2	7.5±0.2	32	8.5±0.5	3.2
44	COR1-105	Ac-DPhe(pI)-Arg-DPhe(pI)-DNal(2’)-NH2	6.3±0.1	500	6.6±0.1	250
45	COR1-13	Ac-DPhe(pI)-Arg-DPhe(pI)-Arg-NH2	7.4±0.1	40	8.5±0.1	3.2
46	COR1-32	Ac-DPhe(pI)-Arg-Nal(2’)-DLys-NH2	6.8±0.3	160	8.0±0.3	10
47	COR1-46	Ac-DPhe(pI)-Arg-Nal(2’)-DNal(2’)-NH2	<5.5		6.7±0.3	200
48	COR1-25	Ac-DPhe(pI)-Arg-Nal(2’)-Arg-NH2	7.8±0.1	16	9.0±0.1	1.0
49	RHC1–65	Ac-Phe(pI)-Nal(2’)-Phe(pI)-Arg-NH2	NA		NA

^aThe indicated errors represent the standard error of the mean determined from at least three independent experiments. Antagonist pA₂ values were determined using a Schild analysis and the agonist NDP-MSH. Antagonist Ki values were converted from pA₂ values using the equation antagonist $\text{Ki}={10}^{-\left({\text{pA}}_{\text{2}}\right)}$. The use of <5.5 indicates that no antagonist potency was observed in the highest concentration range assayed (10,000, 5,000, 1,000, and 500 nM). NA indicates that the tetrapeptide was not assayed as an antagonist.

^bThe AGRP(86–132) data used the same cAMP-based AlphaScreen assay herein.⁶² The miniAGRP(87–120) data are from our laboratory using a cAMP-based β-galactosidase reporter gene assay.^{40, 63}

The importance of the basic residues in position 2 and 4 at the mMC4R can be visualized as pie charts comparing the amino acid substitutions within a position with their ranges of pA₂ values (Figure S2). All the tetrapeptides possessing a DNal(2’) substitution in position 2 resulted in mMC4R antagonist pA₂ values of less than 7 (light and dark red), while an Arg substitution in position 2 resulted in eighteen tetrapeptides with pA₂ values greater than 7 (light and dark green). A similar pattern at position 4 is observed, where eight of sixteen compounds with either the Arg or DLys substitution possess pA₂ values greater than 7 (light and dark green), while only two of eight compounds with the DNal residue have a pA₂ between 7 and 7.9 (light green). Another way to visualize the importance of the basic charges is to compare the mMC4R pA₂ values between paired tetrapeptides, in which three of the four amino acids remain constant while the fourth position is varied (Figure 6 and Figure S3). The difference in mMC4R pA₂ values between DNal(2’) and DPhe(pI) remains relatively flat a position 1 (Figure 6), indicating that both substitutions result in similar mMC4R antagonist potency. In contrast, there is a marked trend for increased pA₂ values when Arg is substituted at position 2 compared to DNal(2’), highlighted by compounds 36 (COR2-3) and 48 (COR1-25) in green in Figure 6. Every Arg substituted tetrapeptide at position 2 possesses a higher mMC4R pA₂ value compared the corresponding DNal(2’) substitution, indicating the importance of Arg at this position. A similar trend is observed at position 4 (Figure S3), with DNal(2’) substitutions possessing lower mMC4R pA₂ values than the corresponding DLys and Arg containing tetrapeptides, while no apparent trends were observed at position 3.

Figure 6:

Comparison of mMC4R pA₂ values for paired tetrapeptides. The paired compounds have three residues in common and vary at the indicated position. Compounds 24 (MDE10–28, blue triangle) and 48 (COR1-25, blue square) are highlighted in the position 1 plot. Compounds 36 (COR2-3, green triangle) and 48 (COR1-25, blue square) are highlighted in the position 2 plot.

The most potent tetrapeptide antagonist at the mMC4R, 48, (COR1-25) [Ac-DPhe(pI)-Arg-Nal(2’)-Arg-NH₂], possessed a pA₂ value of 9.0 (Table 3 and Figure 5). Previous work with longer and more structurally complex compounds such as AGRP, AGRP-derived macrocycles, and the synthetic SHU9119 cyclic peptide with the AlphaScreen Assay resulted in similar pA₂ values at the mMC4R (8.7–9.5).^{40–41, 62, 64–65} The mixture based-positional scan approach was therefore able to select tetrapeptide sequences with antagonist potencies at the mMC4R similar to the endogenous antagonist (AGRP) and the synthetically-developed (SHU9119), indicating the utility of this approach in identifying novel ligands with potent functional activity.

Of the remaining twenty-four tetrapeptides, eleven possessing the Ac-Aromatic-Aromatic-Aromatic-Basic-NH₂ substitution pattern did not possess measurable antagonist activity at the highest concentrations assayed (Table 3). Of the five remaining peptides with this pattern, 4 (COR1-29-2, pA₂ = 6.7), 25 (COR1-80, pA₂ = 6.3), 28 (COR1-95, pA₂ = 6.6), 30 (COR1-90, pA₂ = 5.6), and 36 (COR2-3, pA₂ = 5.8), no strong activity trends were apparent with position 1 and 3 substitutions. Within these five, DLys at position 4 resulted in the most potent mMC4R antagonist activity, which can be observed by comparing the paired tetrapeptides at position 4 [Figure S3, the cluster of DLys and Arg compounds without a corresponding DNal(2’) substitution]. By contrast, the eight compounds with the Ac-Aromatic-Basic-Aromatic-Aromatic-NH₂ substitution pattern possessed mMC4R antagonist activity. Six compounds with this scaffold had pA₂ values between 6.3 and 6.7. The remaining two paired tetrapeptides possessed the sequence Ac-Xxx-Arg-Trp-DNal(2’)-NH₂, where Xxx was DNal(2’) [17 (COR1-60, pA₂ = 7.0)] or DPhe(pI) [41 (COR1-85), pA₂ = 7.6]. Although a basic charge in position 2 generally resulted in more potent mMC4R antagonist than a basic charge in position 4, the potency was still decreased compared to tetrapeptides possessing two basic charges.

Although the tetrapeptide sequences presented herein have not previously been identified, the tripeptide sequence Ac-DNal(2’)-Arg-Nal(2’) (MCL0020) was reported to possess 11.63 and 1115 nM binding affinities at the human (h)MC4R and hMC3R, respectively, and was unable to displace radiolabeled NDP-MSH at up to 10 μM concentrations at the hMC1R.⁶⁶ This tripeptide sequence is contained within the tetrapeptides 22 (COR1-29-8), 23 (COR2-33), and 24 (MDE10–28) from the present study, with an additional DLys, DNal(2’) or Arg added at the fourth position, respectively. The increased potency of 22 (COR1-29-8, pA₂ = 7.8) and 24 (MDE10–28, pA₂ = 8.8) compared to 23 (COR2-33, pA₂ = 6.7) indicates the importance of a basic residue in the fourth position for mMC4R antagonist potency and demonstrates that specific residues in this position can modulate antagonist potency over 100-fold range.

A subsequent structure-activity relationship study around the MCL0020 tripeptide, utilizing L and D isomers of Phe, Nal(1’), and Nal(2’) at the first position and L- and DNal(2’) at the third position indicated a series of compounds with nanomolar to micromolar affinities at the hMC4R.⁶⁷ The most potent compounds possessed DNal(2’) at the first position and either L- or DNal(2’) at the third position (IC₅₀ values of 15.4 and 36.5 nM, respectively).⁶⁷ Similar to the MCL0020 tripeptide, the Ac-DNal(2’)-Arg-DNal(2’) sequence was in three tetrapeptides from the present study [13 (COR1-29-5), 14 (COR2-57), and 15 (COR2-93) possessing DLys, DNal(2’), and Arg at the fourth position, respectively]. The basic residues in the fourth position increased antagonist potency approximately 100-fold at the mMC4R [pA₂ values of 8.2 and 8.3 for DLys and Arg, respectively, compared to 6.3 for DNal(2’)] when combined with this tripeptide sequence. These data indicate the importance of the fourth position within the tetrapeptide sequence, which may not be apparent from these published tripeptide sequences.

Melanocortin-3 Receptor Activity:

Although the forty tetrapeptides utilized in this study were selected based on predicted MC4R antagonist pharmacology, structure-activity relationship (SAR) trends for the mMC3R were also observed. At the highest concentration assayed (100 μM), eleven of the forty tetrapeptides produced a partial agonist response, up to 50% of the maximal signal generated by NDP-MSH. Like at the mMC4R, no compound produced an agonist dose-response curve at the mMC3R.

The sixteen tetrapeptides possessing the Ac-Aromatic-Basic-Aromatic-Basic-NH₂ substitution pattern all possessed measurable mMC3R antagonist activity, albeit with decreased potency when compared to the mMC4R (pA₂ ranges of 5.5 to 7.8 versus 7.6 to 9.0, respectively). The most potent MC3R antagonist, 48 (COR1-25, Table 3 and Figure 5, pA₂ = 7.8), was the same tetrapeptide that resulted in the most potent MC4R antagonist. Similar to the mMC4R antagonist data, the importance of a basic residue in positions 2 and 4 can be seen in the pie charts comparing the different substitutions and mMC3R pA₂ values (Figure S4). Compounds containing a DNal(2’) in position 2 all possessed pA₂ values less than 6, and tetrapeptides with DNal(2’) in position 4 all had pA₂ values less than 7. In contrast, seven compounds with Arg in position 2 and either a DLys or Arg in position 4 had pA₂ values between 7.0 and 7.8, demonstrating the most potent mMC3R antagonist tetrapeptides had two basic residues. Comparing paired tetrapeptides at positions 2 and 4 (Figure S5), also indicates the relative importance of a basic charge compared to the aromatic DNal(2’). For compounds with the Ac-Aromatic-Basic-Aromatic-Basic-NH₂ pattern, there was a general correlation between mMC3R and mMC4R antagonist potencies (Figure S6).

Of the sixteen compounds possessing the Ac-Aromatic-Aromatic-Aromatic-Basic-NH₂ substitution, fourteen did not produce measurable antagonist activity at the mMC3R at the concentrations assayed. The two remaining tetrapeptides from this pattern, 4 (COR1-29-2, pA₂ = 5.8) and 28 (COR1-95, pA₂ = 5.7), possessed measurable mMC4R antagonist activity and both contained DNal(2’) in position 2, Trp in position 3, and DLys in position 4. The eight remaining compounds possessed an Ac-Aromatic-Basic-Aromatic-Aromatic-NH₂ substitution pattern. Two were inactive at the concentrations assayed at the mMC3R [the paired tetrapeptides 23 (COR2-33) and 47 (COR1-46)]. The remaining six possessed pA₂ values between 5.7 to 6.3.

All tetrapeptides with measurable antagonist activity were more potent at the mMC4R compared to the mMC3R, indicating that no ligands were selective for mMC3R antagonism. While the majority of the ligands were less than 100-fold selective for the mMC4R over the mMC3R, three compounds were more than 150-fold selective for the mMC4R [13 (COR1-29-5), 15 (COR2-93), and 16 (COR1-29-6)]. Whereas AGRP-derived macrocyclic octapeptide antagonists have been described as 600- to 800-fold more selective for the mMC4R over the mMC3R,⁴¹ the relative facile synthesis of linear tetrapeptides versus head-to-tail cyclized macrocyclic octapeptides may afford a simpler scaffold to perform SAR campaigns, generating more potent and/or selective ligands and probe molecules. A potential method to identify more selective compounds would be to perform the initial mixture screen at both the MC3R and MC4R, and select ligands for individual compound synthesis based upon potential putative potency differences.

Certain residues appeared to drive the mMC4R over the mMC3R. After converting pA₂ to antagonist Ki values [antagonist Ki = 10^−(pA2)], antagonist Ki values could be compared between paired compounds at the mMC4R and mMC3R (Figure S7). Neither the DNal(2’) or DPhe(pI) substitution at position 1 resulted in a uniform trend. At position 2, every Arg substituted tetrapeptide possessed increased mMC4R selectivity compared to DNal(2’). While there is not an overall trend between all the substitutions at position 3, mMC4R selectivity was uniformly increased when DNal(2’) was incorporated versus DPhe(pI). Basic residues in position 4 also trended towards increased mMC4R selectivity versus DNal(2’)-substituted ligands. These data suggest certain residues can be incorporated to increase mMC4R antagonist selectivity over the mMC3R within the tetrapeptide scaffold.

Melanocortin-1 Receptor Activity:

Like the mMC3R data, SAR information can be obtained from the mMC1R data, although for this receptor a different substitution pattern was found to be important for activity. Of the forty tetrapeptides assayed at the mMC1R, six possessed full agonist activity [19 (COR1-29-7), 20 (COR2-15), 21 (COR2-105), 43 (COR1-100), 44 (COR1-105), and 45 (COR1-13)]. The full agonist efficacy of 45 (COR1-13) is presented in Figure 7. Common to each of these tetrapeptides is an Ac-Xxx-Arg-DPhe(pI)-Yyy-NH₂ motif that inverts the natural Phe-Arg position of the endogenous agonist sequence His-Phe-Arg-Trp. The Ac-Xxx-Arg-DPhe(pI)-Yyy-NH₂ motif has previously been described as possessing full agonist activity at the mMC1R.^{46–47, 68} In particular, the pharmacology of 20 (COR2-15) has previously been reported as an agonist at the mMC1R (EC₅₀ = 4,000 nM) with partial agonist efficacy at the mMC3R, mMC4R, and mMC5R (40%, 37%, and 43% of the maximal NDP-MSH signal, respectively), and as an antagonist at the mMC3R (pA₂ = 5.6) and mMC4R (pA₂ = 5.9).⁴⁶ These values are similar to the agonist activity at the mMC1R (EC₅₀ = 5,000 nM), partial agonist efficacy at the mMC3R, mMC4R, and mMC5R (40%, 30%, and 40% of the maximal NDP-MSH signal, respectively), and antagonist activity at the mMC3R (pA₂ = 6.1) and mMC4R (pA₂ = 6.5) reported herein [Tables 2 and 3, 20 (COR2-15)]. Within this set of full mMC1R agonist tetrapeptides, a basic amino acid (DLys or Arg) at position 4 increased agonist potency more than 10-fold compared to an aromatic [DNal(2’)] residue.

Figure 7:

Illustration of the agonist pharmacology of NDP-MSH, 9 (COR2-87), 42 (COR1-7), 44 (COR1-105), 45 (COR1-13), and 48 (COR1-25) at the mMC1R and mMC5R.

All the full agonists at the mMC1R possessed a DPhe(pI) group in the third position. Select tetrapeptides containing other aromatic amino acids, including DNal(2’) [13 (COR1-29-5), 14 (COR2-57), 15 (COR2-93), 37 (COR2-111), and 39 (COR1-19)], Nal(2’) [22 (COR1-29-8)], or Trp [42 (COR1-7, Figure 7)], in the third position resulted in partial efficacy agonists at the mMC1R (30–85% maximal NDP-MSH signal, EC₅₀ values 400–7,900 nM). Furthermore, while five of the six tetrapeptides with the Ac-Xxx-Arg-DNal(2’)-Yyy-NH₂ scaffold resulted in agonists with partial efficacy at the mMC1R, only one of the six tetrapeptides with either the Ac-Xxx-Arg-Nal(2’)-Yyy-NH₂ or the Ac-Xxx-Arg-Trp-Yyy-NH₂ sequences generated a partial mMC1R agonist response. The sigmoidal dose-response curve observed for eleven of the twelve tetrapeptides with a D-aromatic residue [DPhe(pI) or DNal(2’)] in the fourth position indicates that both the stereochemistry and functional group may be important for activation of the mMC1R.

Melanocortin-5 Receptor Activity:

While the tetrapeptides were not selected for mMC5R activity, select SAR trends were observed at this receptor. The sixteen compounds with the Ac-Aromatic-Basic-Aromatic-Basic-NH₂ substitution pattern, important for mMC3R and mMC4R antagonist activity, did not possess mMC5R agonist potency at up to 100 μM concentrations. Of the eight compounds with the Ac-Aromatic-Basic-Aromatic-Aromatic-NH₂ motif, three were inactive. The other five [14 (COR2-57), 20 (COR2-15), 23 (COR2-33), 38 (COR2-27), and 44 (COR1-105, Figure 7)] were able to partially stimulate the mMC5R at 100 μM concentrations (20% to 65% of the maximal NDP-MSH signal, Table 2). The sixteen remaining compounds had the Ac-Aromatic-Aromatic-Aromatic-Basic-NH₂ pattern. Only one of the eight compounds in this set with DLys at position 4 [34, (COR2-117)] resulted in partial activation of the mMC5R at 100 μM (50% of the maximal NDP-MSH signal). The other seven tetrapeptides were inactive at the highest assayed concentrations. Of the eight Ac-Aromatic-Aromatic-Aromatic-Basic-NH₂ tetrapeptides with Arg in position 4, one was inactive [30 (COR1-90)] and two [6 (MDE10–29) and 12 (MDE10–63)] partially stimulated the mMC5R at the highest concentrations assayed (45% and 40% of the maximal NDP-MSH signal, respectively). The remaining five compounds [3 (COR2-99), 9 (COR2-87), 27 (COR1-65), 33 (COR1-55), and 36 (COR2-3)] were full mMC5R agonists with sub-micromolar potencies (EC₅₀ between 30 and 800 nM). The nanomolar mMC5R potency of 9 (COR2-87) can be observed in Figure 7. These compounds either partially activated or were inactive as agonists at the highest assayed concentrations at the mMC1R, mMC3R, and mMC4R, and did not act as antagonists at the mMC3R. Only one of these compounds, 36 (COR2-3), possessed mMC4R antagonist activity (pA₂ = 5.8). With the minimal observed activity at the other melanocortin receptors, the Ac-Aromatic-Aromatic-Aromatic-Arg-NH₂ motif may be further optimized for mMC5R agonist activity.

As an initial approach to identify the key residues for mMC5R selective agonism, an L-residue scan was performed on the 33 (COR1-55) tetrapeptide [Ac-DPhe(pI)-DNal(2’)-DPhe(pI)-Arg-NH₂] in which the D amino acids in the first three positions were systematically replaced with the L-isomers. While 33 (COR1-55) compound containing three D-aromatic residues was soluble and could be purified, the six tetrapeptides containing one or two D-amino acids were not soluble and could not be purified. The all-L tetrapeptide [49 (RHC1–65), Ac-Phe(pI)-Nal(2’)-Phe(pI)-Arg-NH₂] was able to partially stimulate the mMC1R and was inactive at the mMC3R and mMC4R up to 100 μM concentrations, similar to 33 (COR1-55), but did not possess agonist activity at the mMC5R (Table 2). These data suggest that for this analog, the D-stereochemistry in the first three positions is preferred over the L-sterochemistry for the observed mMC5R-selective agonism, which may be utilized in the development of future tetrapeptide probes.

Potent MC5R agonists have been reported in the past. The mixed pharmacology of SHU9119 (Ac-Nle-c[Asp-His-DNal(2’)-Arg-Trp-Lys]-NH₂; MC1R/MC5R agonist, MC3R/MC4R antagonist) was first reported in 1995.³⁷ Replacing the His within the SHU9119 scaffold with Pro, resulting in the cyclic peptide Ac-Nle-c[Asp-Pro-DNal(2’)-Arg-Trp-Lys]-NH₂ (PG901), was reported to possess sub-nanomolar agonist potency at the hMC5R.^69–70 Similar to SHU9119, PG901 was reported to possess nanomolar potent antagonist activity at the hMC3R and hMC4R (pA₂ = 9.0 and 9.3, respectively).^69–70 While activity at the MC1R was not reported, due to the similar structure of PG901 to SHU9119 and similar pharmacology at the hMC3R, hMC4R, and hMC5R, PG901 may also have potent agonist activity at the MC1R. Potential advantages to the tetrapeptides reported herein may be synthetic tractability due to the lack of the lactam bridge, and greater selectivity for the MC5R over the other melanocortin receptors. Dosing of these compounds may be beneficial since they do not possess antagonist activities at the MC3R and MC4R, and may more clearly elucidate the in vivo functions of the MC5R, albeit with decreased potency compared to PG901.

Substitutions of the linear NDP-MSH have also been reported to result in selective and potent MC5R agonists. A sub-nanomolar MC5R agonist was identified (Ac-Ser-Tyr-Ser-Nle-Glu-Oic-DBip-Pip-Trp-Gly-Lys-Pro-Val-NH₂) that minimally activated the other melanocortin receptors at 5 μM, and possessed greater than 3200-fold selectivity for the MC5R as assessed by radiolabeled ¹²⁵I-NDP-MSH displacement.⁷¹ Truncation to Ac-Oic-DBip-Pip-Trp-NH₂ resulted in micromolar agonist potency at the MC5R (1.2 μM).⁷¹ A similar substitution pattern of the His-DNal(2’)-Arg tripeptide sequence in SHU9119 with Oic-DBip-Pip, resulting in Ac-Nle-c[Asp-Oic-DBip-Pip-Trp-Lys]-NH₂, also resulted in potent and selective MC5R agonism.⁷² This compound minimally activated the MC1R, MC3R, and MC4R (up to 40% at 2 μM) while activating the MC5R at a 0.99 nM concentration.⁷² Additionally, this compound displaced radiolabeled NDP-MSH at the MC5R at a 0.95 nM concentration, but was unable to displace NDP-MSH at up to 5,000 nM concentrations at the MC1R, MC3R, and MC4R.⁷² These results indicate that inserting an active MC5R tetrapeptide sequence into a linear or cyclically constrained scaffold can increase potency and/or selectivity. The most potent tetrapeptides from the present study at the MC5R [9 (COR2-87), 3 (COR2-99), and 33 (COR1-55); EC₅₀ = 30, 130, and 130 nM, respectively] were more potent than the equivalent Ac-Oic-DBip-Pip-Trp-NH₂ sequence (1.2 μM), suggesting that insertion of this new tetrapeptide into the NDP-MSH or MTII/SHU9119 scaffold may also increase MC5R agonist potency/selectivity.

Conclusions

To identify novel MC4R antagonist sequences, an unbiased functional mixture-based positional scanning approach was utilized in an antagonist assay paradigm. Initially, mixtures comprising a library of 12,960,000 compounds were screened in the absence and presence of 0.5 nM NDP-MSH at a 50 μg/mL concentration. Twenty-six mixtures were then screened using two antagonist concentrations (25 and 50 μg/mL) with NDP-MSH (0.5 nM). Following library deconvolution, the defined amino acids of the most active mixtures at each position were selected for generating a library of forty-eight individual compounds. From this library, eight compounds (consisting of four aromatic amino acids, an acetylated N-terminal, and an amidated C-terminal) were not soluble in the purification solvents and were not advanced. The remaining forty compounds were screened for agonist activity at the mMC1R, mMC3R, mMC4R, and mMC5R. Compounds that did not stimulate the mMC3R and mMC4R were then screened as antagonists at these receptors with NDP-MSH serving as the agonist. Thirteen nanomolar potent mMC4R antagonist tetrapeptides were identified, with a general sequence of Ac-Aromatic-Basic-Aromatic-Basic-NH₂. One tetrapeptide (48, COR1-25) was identified that is ca equipotent to the endogenous AGRP(86–132) antagonist (consisting of 46 amino acids and 5 disulfide bridges) at the MC4R, providing a potential new molecular probe for in vitro and in vivo studies. Three compounds possessed greater than 100-fold selectivity for the mMC4R over the mMC3R. A set of six full mMC1R agonists were identified with an inverted Arg-Phe motif. Five sub-micromolar mMC5R agonist tetrapeptides were also reported with minimal activation of the remaining receptors. Compounds from this library may prove useful in generating novel probes and therapeutic leads for the treatment of disease states of negative energy balance, including cachexia associated with cancer.

Experimental

Mixture-Based Positional Scanning Library:

As previously described,^{45–46, 52, 73–74} the TPI924 library was synthesized using the simultaneous multiple-peptide synthesis approach. This TPI924 Ac-tetrapeptide-NH₂ template [Ac-R1-R2-R3-R4-NH₂] library contains 12,960,000 tetrapeptides in 240 mixtures composed of 60 amino acid building block combinations at each of the “R” positions. The sixty amino acids used for each of the four “R” positions within the tetrapeptide template include Ala, Asp, Glu, Phe, Gly, His, Ile, Lys, Leu, Met, Asn, Pro, Gln, Arg, Ser, Thr, Val, Trp, Tyr, DAla, DAsp, DGlu, DPhe, DHis, DIle, DLys, DLeu, DMet, DAsn, DPro, DGln, DArg, DSer, DThr, DVal, DTrp, DTyr, Nle, DNle, Cha, DCha, Ala(3-Pyr), DAla(3-Pyr), Ala(2-Thi), DAla(2-Thi), Tic, DTic, Phe(pCl), DPhe(pCl), Phe(pI), DPhe(pI), Phe(pNO2), DPhe(pNO2), Nal(2’), DNal(2’), β-Ala, ε-aminocaproic acid, Met[O2], dehydPro, and Tyr(3-I).

Fluorescent β-Galactosidase Assay:

A fluorescent β-galactosidase bioassay was used for the primary screen of the mixture-based positional scanning library TPI924, based upon an absorbance β-galactosidase bioassay previously described.⁵³ HEK-293 cells stably expressing the mMC4R were transfected with 4 μg of CRE/β-galactosidase using the calcium phosphate method.⁷⁵ After 24 h, cells were plated onto collagen-treated black 96-well plates (Corning) and incubated at 37 °C with 5% CO₂. At 48 h post-transfection, the cell media was aspirated and the cells were stimulated with compounds mixtures, in both agonist and antagonist experimental paradigms. Compound mixtures were dissolved 1:1 DMF:H₂O at 20 mg/mL stock concentrations, diluted with H₂O to 2 mg/mL working concentrations, and stored at −20 °C until use. For fluorescence agonist experiments, 40 μL of the TPI924 peptide mixture (50 μg/mL) in assay media (DMEM containing 0.1 mg/mL BSA and 0.1 mM isobutylmethylxanthine) was added to each well. For fluorescence antagonist studies, 40 μL of the TPI mixture (50 μg/mL) and NDP-MSH (0.5 nM) in assay media was added to each well. Triplicate wells were used for each experimental mixture, and at least three independent experiments were performed. Controls included NDP-MSH (10⁻⁶ to 10⁻¹³ M), forskolin (100 nM), and assay media. The plates were incubated at 37 °C with 5% CO₂ for 6 h. Post-stimulation, the media was aspirated and 120 μL of lysis buffer (7:1 Z-buffer [60 mM Na₂PO₄·7H₂O, 40 mM Na₂H₂PO₄·H₂O, 10 mM KCl, 1 mM MgSO₄·7H₂O, pH = 7.0]:1% Triton X-100 in H₂O [v/v]) was added. Plates were stored at −80°C for up to one week.

The plates were thawed and 10 μL aliquots of cell lysate were added to 200 μL of BioRad dye solution (1:4 dye solution:H₂O) in another clear 96-well plate. Absorbance was read using a FlexStation 3 (Molecular Devices) at λ = 595 nm. To the remaining cell lysate, 30 μL of substrate buffer [0.1 mM 4-methylumbelliferyl-β-D-galactopyranoside (4-MUG), 0.045 M β-mercaptoethanol, dissolved in Z-buffer] was added to each well, and the plates incubated at 37 °C. One plate was monitored in fluorescence mode (excitation λ = 350 nm, emission λ = 450 nm), until the positive controls reached 8,000 RFU, at which point 75 μL of the stop buffer (300 mM glycine, 15 mM EDTA, pH 11.2) was added to each well. Fluorescence readings were measured on a FlexStation 3 (Molecular Devices) in Endpoint Fluorescence mode (excitation λ = 350 nm, emission λ = 450 nm). For each well, the plate was normalized to the corresponding well protein level, plate basal signal, and the maximal NDP-MSH signal (10⁻⁶ to 10⁻⁸ M). For agonists screens, the normalized well signal was converted to a percentage from plate basal signal to the plate maximal NDP-MSH response. In the initial antagonist screen, the normalized well signal was converted to a percentage from the median plate experimental signal to basal activity. In the two-point follow-up antagonist screen, the normalized well signal was converted to a percentage from plate 0.5 nM NDP-MSH signal to basal activity. Each mixture was screened using triplicate wells in at least two independent experiments. Agonist and antagonist means were calculated from each mixture replicate assays. Singleton outlying points were eliminated using the quartile method.⁷⁶

Tetrapeptide Synthesis and Purification:

The amino acids Fmoc-Arg(Pbf), Fmoc-Trp(Boc), and Fmoc-DLys(Boc), the Rink-amide MBHA resin, and coupling reagent 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate (HBTU) were purchased from Peptides International (Louisville, KY). The Fmoc-DNal(2’) amino acid was purchased from Bachem and Peptides International. The Fmoc-Nal(2’) residue was purchased from SyntheTech. The Fmoc-DPhe(pI) amino acid was purchased from Alfa Aesar. Dichloromethane (DCM), methanol (MeOH), acetonitrile (ACN), dimethylformamide (DMF), and anhydrous diethyl ether were purchased from Fisher (Fair Lawn, NJ). Trifluoroacetic acid (TFA), dimethyl sulfoxide (DMSO), piperidine, triisopropylsilane (TIS), thioanisole, and N,N-diisopropylethylamine (DIEA) were purchased from Sigma-Aldrich (St. Louis, MO). All reagents and chemicals were ACS grade or better and were used without further purification.

The tetrapeptides identified from the library deconvolution were synthesized using standard N-α-fluorenylmethoxycarbonyl (Fmoc) methodologies,^55–56 individually on a microwave synthesizer (Discover SPS; CEM, Matthews, NC) or in parallel (LabTech I; Advanced ChemTech, Louisville, KY). The resin was allowed to swell in DCM before iterative Fmoc deprotection and amino acid coupling steps were used to assemble the desired tetrapeptides. For peptides synthesized on the microwave synthesizer, the Fmoc group was removed using a two-step deprotection strategy with 20% piperidine in DMF (1 × 2 min at room temperature, followed by 1 × 4 min using microwave irradiation to 75 °C, and 30W). Microwave coupling reactions were carried out at 75 °C and 30W for 5 min (10 min for coupling Arg residues). For peptides synthesized in a parallel, a two-step deprotection strategy was used (1 × 5 min, 1 × 15 min at room temperature) with 20% piperidine in DMF. Parallel coupling reactions were carried out at room temperature for 45 min. For most coupling reactions, 3.1 eq of the amino acid, 3 eq HBTU, and 5 eq of DIEA were used. For Arg coupling, high equivalents were employed (5.1 eq Arg, 5 eq HBTU, and 7.1 eq DIEA). In all cases, reaction progress was monitored using a ninhydrin assay,⁷⁷ and deprotection and coupling reactions were repeated if necessary. Following the removal of the terminal Fmoc group, tetrapeptides were acetylated by adding a 3:1 acetic anhydride:pyridine solution and mixing for 30 min at room temperature. The resin was then washed with DMF, DCM, and MeOH. Tetrapeptides were cleaved from resin and side-chain deprotected using a 91:3:3:3 mixture of TFA:H₂O:TIS:thioanisole for 2 h at room temperature, and then precipitated using ice-cold diethyl ether. Crude peptides were pelleted on a Sorvall Legend XTR centrifuge, and dried overnight in a vacuum desiccator.

The crude tetrapeptides were purified on a C18 RP-HPLC semipreparative column (Vydac 218TP1010, 1.0 cm × 25 cm) using a Shimadzu system equipped with a UV detector. Peptides are ≥95% pure as ascertained by analytical RP-HPLC (Vydac 218TP104, 0.46 cm × 25 cm) on a Shimadzu system equipped with a PDA detector in two diverse solvent systems (methanol and acetonitrile). The peptides possessed the correct average molecule mass by ESI-MS (Bruker BioTOF II ESI/TOF-MS; LeClaire-Dow Instrumentation Facility, University of Minnesota).

cAMP AlphaScreen Bioassay:

The purified tetrapeptides were dissolved in DMSO at a stock concentration of 10⁻² M (NDP-MSH in H₂O at a stock concentration of 10⁻⁴ M), and assayed using HEK293 cells stably expressing the mouse MC1R, MC3R, MC4R, and MC5R using the AlphaScreen cAMP bioassay (PerkinElmer) according to the manufacturer’s instructions and as previously described.^{57–58, 60}

Briefly, cells 70–90% confluent were dislodged with Versene (Gibco) at 37 °C and plated 10,000 cells/well in a 384-well plate (Optiplate) with 10 μL of freshly prepared stimulation buffer (1 × HBSS, 5 mM HEPES, 0.5 mM IBMX, 0.1% BSA, pH = 7.4) with 0.5 μg of anti-cAMP acceptor beads per well. The cells were stimulated with the addition of 5 μL of stimulation buffer containing peptide (concentrations from 10–4 to 10–13 M, determined by ligand potency) or forskolin (10–4 M) and incubated in the dark at room temperature for 2 h.

Following stimulation, streptavidin donor beads (0.5 μg) and biotinylated-cAMP (0.62 μmol) were added to the wells in a green light environment with 10 μL of lysis buffer (5 mM HEPES, 0.3% Tween-20, 0.1% BSA, pH = 7.4) and the plates were incubated in the dark at room temperature for an additional 2 h. Plates were read on a Enspire (PerkinElmer) Alpha-plate reader using a pre-normalized assay protocol (set by the manufacturer).

Data Analysis:

The pA₂ and EC₅₀ values represent the mean of at least three independent experiments performed in duplicate replicates. Compounds that were not active in two independent experiments at the concentrations assayed (>100 μM for agonist assays; 10,000, 5,000, 1,000, and 500 nM for antagonist assays) were not furthered examined. The pA₂ and EC₅₀ estimates, and associated standard errors (SEM) were determined by fitting the data to a nonlinear least-squares analysis using the PRISM program (version 4.0, GraphPad Inc). The peptides were assayed as TFA salts and not corrected for peptide content.

Abbreviations Used:

ACN

acetonitrile

ACTH

adrenocorticotropic hormone

AGRP

agouti-related protein

ASP

agouti-signaling protein

cAMP

3’,5’-cyclic adenosine monophosphate

DIEA

diisopropylamine

DMEM

Dulbecco’s modified eagle medium

ESI-MS

electrospray ionization mass spectrometry

FDA

U.S. Food and Drug Administration

Fmoc

fluorenylmethoxycarbonyl

HBSS

Hanks’ balanced salt solution

HBTU

2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate

IBMX

isobutylmethylxanthine

LEPR

leptin receptor

LY2112688

Ac-D-Arg-c-[Cys-Glu-His-D-Phe-Arg-Trp-Cys]-NH₂

MBHA

methylbenzhydrylamine

MC1R

melanocortin 1 receptor

MC2R

melanocortin 2 receptor

MC3R

melanocortin 3 receptor

MC4R

melanocortin 4 receptor

MC5R

melanocortin 5 receptor

MCR

melanocortin receptor

MeOH

methanol

MSH

melanocyte-stimulating hormone

MTII

melanotan II

NDP-MSH

4-norleucine, 7-D-phenylalanine-α-melanocyte-stimulating hormone (melanotan I)

PCSK1

proprotein convertase subtilisin/kexin type 1

PDA

photodiode-array detector

POMC

pro-opiomelanocortin

RP-HPLC

reversed-phase high-pressure liquid chromatography

SAR

structure-activity relationship

SEM

standard error of the mean

t-Boc

tert-butyloxycarbonyl

TIS

triisopropylsilane