The Multifaceted Melanocortin Receptors

Abstract

The 5 known melanocortin receptors (MCs) have established physiological roles. With the exception of MC₂, these receptors can behave unpredictably, and since they are more widely expressed than their established roles would suggest, it is likely that they have other poorly characterized functions. The aim of this review is to discuss some of the less well-explored aspects of the 4 enigmatic members of this receptor family (MC_1,3-5) and describe how these are multifaceted G protein–coupled receptors (GPCRs). These receptors appear to be promiscuous in that they bind several endogenous agonists (products of the proopiomelanocortin [POMC] gene) and antagonists but with inconsistent relative affinities and effects. We propose that this is a result of posttranslational modifications that determine receptor localization within nanodomains. Within each nanodomain there will be a variety of proteins, including ion channels, modifying proteins, and other GPCRs, that can interact with the MCs to alter the availability of receptor at the cell surface as well as the intracellular signaling resulting from receptor activation. Different combinations of interacting proteins and MCs may therefore give rise to the complex and inconsistent functional profiles reported for the MCs. For further progress in understanding this family, improved characterization of tissue-specific functions is required. Current evidence for interactions of these receptors with a range of partners, resulting in modulation of cell signaling, suggests that each should be studied within the full context of their interacting partners. The role of physiological status in determining this context also remains to be characterized.

Melanocortin receptors (MCs) are instrumental for a range of clinically relevant physiological functions. MC₁ mediates pigmentation of both skin and hair, MC₂ is required for adrenal steroidogenesis and therefore the stress response, MC₃ and MC₄ modulate the central control of food intake and satiety, and MC₅ regulates sebogenesis. These are essential functions: it might therefore be assumed that the receptors are both structurally and functionally well characterized. The aim of this review is to demonstrate that there are many aspects of the multifaceted MCs that warrant further investigation.

For each of the 5 receptors identified to date, the primary and secondary structures are well described, but the tertiary structures are only recently being revealed for some. Understanding of how structure relates to function is therefore a work in progress. Except for MC₂, the MCs bind multiple ligands, and this lack of specificity is an unusual feature of a G protein–coupled receptor (GPCR). Another exceptional feature is that some MCs also have endogenous antagonists. Receptor activation is associated with a range of cellular responses, which at first was attributed to the multiple ligands that can activate the receptors. However, it has become increasingly apparent that this explanation is inadequate: the reality is far more complex and context dependent. MCs are more widely expressed throughout the body than the functions described in the opening sentences might suggest, albeit in some tissues their expression is very low. Their functions in these other tissues are not well characterized if indeed known at all. More than one MC type may be expressed in a single tissue and even within the same cell. In vitro data suggest that MCs can form heterodimers that may affect signaling on activation. Two melanocortin receptor accessory proteins, the MRAPs, interact with the MCs to influence MC signaling. MC signaling can be further modified not only by the interactions of the MRAPs and MCs with each other but also through specific interactions with certain other proteins as well as other GPCRs. Hence, multiple factors need to be considered when trying to characterize each of the MCs before we can further our understanding of the multifaceted MC family.

This review will not consider MC₂: it is the “black sheep” of the family in that it only binds one of the melanocortin peptides. In the future though, we may learn more about the other MCs by exploring why MC₂ is different.

Established Physiological Roles of MCs and Consequences of Genetic Variation

The complexity of the MCs and their multifaceted features belies textbook views of a simpler range of functions, many of which are underpinned by overt human and mouse phenotypes resulting from mutations: these are described briefly below to put in context the more complex aspects that we will describe later.

MC₁

In epidermal and hair follicle melanocytes, MC₁ regulates the synthesis of eumelanin (black/brown) pigments. The MC₁ gene is highly polymorphic in individuals of European ancestry, but not in those of African ancestry, and many of the 80 plus variants identified to date produce a nonfunctional receptor (1). Loss of function in MC₁ results in an increase in the relative amount of phaeomelanin (yellow/red) to eumelanin synthesized. The resulting phenotype is fair skin, freckles and red hair (red hair color variants). An association between fair skin and the incidence of melanoma has generated interest in these variants. Red hair color variants are associated with an increased susceptibility to developing both melanoma and nonmelanoma skin cancers; however, not all variants associated with an increase in skin cancer susceptibility are also associated with changes in pigmentation (2). Not all loss of function is associated with reduced cyclic adenosine monophosphate (cAMP) activity on receptor activation as some of the variants result in a reduced number of receptors at the cell membrane suggesting dysfunctional receptor trafficking (3).

MC₃

The MC₃ knockout (KO) mouse has reduced lean body mass and increased fat mass resulting in an obese phenotype (4, 5). The association between human MC₃ gene variants and obesity is still unclear due to the rarity of such variants (6). The 2 most common variants, T6K and V81I, have been reported by some as associated with an obese phenotype but by others not: a mouse model that has these 2 variants is obese (6).

Both male and female MC₃ KO mice have impaired linear growth (5). Screening whole-exome sequence data of 200 000 individuals from the UK Biobank revealed more than 170 different variants in the human MC₃ gene (7): all are exceptionally rare. The researchers selected the 3 most common variants and performed sophisticated analyses using all 500 000 participants in the UK Biobank to demonstrate that these 3 variants were each associated with shorter stature (7). MC₃ colocalizes to growth hormone–releasing hormone neurons in the hypothalamus, and the authors suggest that MC₃ may therefore act at the level of the hypothalamus to regulate height. Rat anterior pituitary somatotrophs both express MC₃ and respond to melanocortins (8, 9) so the potential involvement of pituitary function in this phenotype should not be ignored.

MC₄

Like the MC₃ KO, the MC₄ KO mouse is obese, however, there are substantial differences between the 2 KOs; in particular, the MC₃ KO is hypophagic and has reduced linear growth while the MC₄ KO mouse is hyperphagic with increased linear growth (10). Appreciation of a possible role for MC₄ in regulating body weight in the mouse (10, 11) prompted a search for variants resulting in obesity in humans. Back-to-back publications reported the identification of 2 individuals and some of their family members who were heterozygous for a rare frame-shift variant that resulted in a truncated MC₄ and therefore a nonfunctional receptor: the affected individuals were all obese (12, 13). Since then, the focus on MC₄ has been mainly on its roles in regulating appetite. Using publicly available data, a number of heterozygous loss-of-function variants in MC₄ with associations to body weight have been identified (14, 15). Variants with a loss of function resulting in reduced generation of cAMP on receptor activation are associated with weight gain, while gain-of-function mutations, which result in biased increased beta-arrestin recruitment followed by increased mitogen-activated protein kinase (MAPK) pathway activation, are associated with a lean phenotype (14). Body weight is not just a function of appetite: evidence is accumulating that MC₄ in the dorsal raphe may also have a role in regulating both thermogenesis and locomotion, and hence energy expenditure (16). Both proopiomelanocortin (POMC) and agouti-related protein (AgRP) neurons in the arcuate have projections to the dorsal raphe (17, 18). Within the dorsal raphe, there are both GABAmergic and glutamergic neurons that express MC₄ (18, 19). Activation of GABAergic neurons by α-melanocyte stimulating hormone (α-MSH) results in decreases in firing rate concomitant with decreases in food intake (19, 20). Increased prolylcarboxypeptidase, an enzyme that results in decreases in available synaptic α-MSH (21), was associated with increased thermogenesis and locomotion (20). Arcuate AgRP acts as an inverse agonist on MC₄ expressing-glutamatergic neurons within the dorsal raphe, resulting in activation of a cluster of 5HT-neurons also within the dorsal raphe. These serotonergic neurons stimulate thermogenesis without eliciting an effect on food intake (18).

To date the focus has remained on centrally expressed MC₄ and its role in body weight regulation; however, there is evidence for MC₄ expression in the periphery (EMBL-EBI gene expression atlas).

MC₅

The only phenotype observed in global MC₅ KOs was that the mice took longer to dry their fur after doing swim tests because of reduced sebogenesis (22). In a study of the human gene, 5 variants were identified in a small sample of individuals with skin/sebaceous gland disorders, however, these same variants were also found with a similar distribution in individuals of a wide range of ethnicities who were phenotypically normal (23). In mouse models, there are also reports of roles for MC₅ in regulating immunological responses in autoimmune disorders of the eye, fatty acid oxidation in skeletal muscle, and lipolysis in adipocytes (24-28). To date there are no reported variants in the human gene associated with any of these roles.

Tissue Distribution of the MCs

Perhaps based on the results of the studies described above, it is generally thought that MC₁ is confined to the integumentary system, MC₃ and MC₄ to the central nervous system (CNS), and MC₅ to exocrine glands. To date, the validity of these conclusions has been hampered by the inability to specifically identify the different MCs using immunohistochemical approaches. The commercially available antibodies for the MCs are not specific (example (29)). Databases such as the EMBL-EBI gene expression atlas suggest that all 4 receptors are more widely distributed throughout the body and basic searches of available literature identifies multiple reports of expression in other tissues, albeit with varying strength of evidence. Importantly, it is evident that some tissues express more than one type of MC, perhaps even within the same cell (30). In vitro, it is known that MCs can heterodimerize with each other, so a better understanding of within-tissue expression is required. The advent of multiplex nucleic acid in situ hybridization technologies, like RNAscope are enabling better precision in identifying MC expression patterns.

The lack of clarity of the tissue distribution of the MCs has important consequences for fully understanding the etiology of some of the phenotypes associated with MC variants, as these may in part be due to dysfunction of the receptor in peripheral tissues. For example, MC₄ is expressed in the heart (EMBL-EBI gene expression atlas (31),) and therefore some of the associations with cardiovascular dysfunction (32, 33) may be due to direct effects on cardiac function and not sequelae of obesity and/or central MC₄ effects. Insulin release is decreased in both lean and obese rats following treatment with NDP-MSH, a synthetic agonist of MC₄ (34). Given that MC₄ is expressed in the pancreas (EMBL-EBI gene expression atlas (34)), dysfunctional insulin release from the pancreas may contribute to the obesity linked to MC₄.

Why are the MCs Unique Among GPCRs?

MC Ligands

Except for MC₂, which is highly selective for adrenocorticotrophic hormone (ACTH), the other MCs, MC_1,3-5, interact with each of the melanocortin proteins derived from the posttranslational cleavage products of the proopiomelanocortin (POMC) gene. The melanocortin proteins are alpha-, beta-, and gamma-melanocyte stimulating hormone (α-, β-, γ-MSH) and ACTH. All the melanocortin ligands, have a conserved HFRW motif (35) with the motif found at the base of the “U” in their U-shaped three-dimensional structures. The benzene ring of the phenylalanine of the HFRW motif penetrates deeply into the transmembrane domain (TMD) core of the receptor (36-38) and results in the downward movement of 2 phenylalanines (F257 and F280) in MC₁ and a leucine (L133) in MC₄. The downward movement of these residues in turn pushes on residues (W254 on MC₁ and W258 on MC₄) that act as toggle switches on TMD6. When switched on, TMD6 moves outward and the receptors are activated.

There are also 2 other gene products that bind to MCs: an inverse agonist, agouti-related protein (AgRP), which is specific for MC₃ and MC₄ (39, 40); and an antagonist of α-MSH, agouti-signaling peptide (ASIP), which competes for binding to MC₁ and MC₄ (41). The ligands can be released into the circulation or act in an autocrine or paracrine way. This ligand diversity and the inconsistent potencies at each MC (described further below) is unique among GPCRs.

MC Structure

The 5 MCs identified to date are all members of the α-subfamily of class A (rhodopsin-like) GPCRs and the human receptors share 42% to 67% of their amino acid sequences (42). There are strong similarities between the reported tertiary structures of MC₁ and MC₄ (36-38); hence, it could be assumed that the other MCs are also structurally similar.

The MCs have several structural features that set them apart from other class A GPCRs. First, the receptors are short (ranging from 297 to 360 amino acids) and compared with other class A GPCRs, they have relatively short N- and C- termini (42). Secondly, both MC₁ and MC₄ have a wide extracellular opening to the orthosteric ligand binding pocket (36-38). The width is due to an exceptionally short second extracellular loop (ECL2), a lack of the conserved cysteines in TMD3 and ECL2 found in other class A GPCRs, and the absence of conserved prolines in TMD2 and TMD5 that are present in other class A GPCRs (42). Extracellular Ca²⁺ has long been recognized as a cofactor for melanocortin binding (43). Within TMD2 and TMD3 are 3 conserved residues unique to the MCs, which form a Ca²⁺-binding pocket in conjunction with 3 conserved amino acids in the ligands (36-38). Calcium ion binding is important for agonist interaction but not for that of antagonist (37). To date, no one has reported the tertiary structure in the presence of the MRAPs and/or other GPCRs that the MCs are known to interact with.

Posttranslational Modification of MCs

Several studies have demonstrated the importance of the conserved cysteine(s) in the C-terminus for normal function of the MCs. These cysteines are sites for posttranslational modification by palmitoylation, which involves the enzymatic addition and removal of a palmitic acid to the cysteine.

Lack of palmitoylation of C315 in the cytoplasmic tail of MC₁ prevents proper receptor function. The zDHHC-protein acyl transferase (zDHHC-PAT) zDHHC-PAT13, responsible for palmitoylation of MC₁, is phosphorylated by UVB light (44). Increasing the interaction of MC₁ with phosphorylated zDHHC-PAT13, results in greater MC₁ activation as seen by increases in palmitoylation, cAMP production, and DNA repair and decreases in cell senescence (44). In the presence of a mutated palmitoylation site, no rescue was achieved with increased phosphorylation and/or increased amounts of ZDHHC-PAT13 (44).

In humans, 2 cysteine residues in the cytoplasmic tail of MC₄ have been identified as predicted sites for palmitoylation (https://swisspalm.org/), which has been confirmed in studies by Moore and Mirshahi (45). This group has also suggested a functional consequence of loss of palmitoylation. MC₄ variants that result in a truncation of the region of the cytoplasmic tail that is palmitoylated leads to loss of receptor function and are associated with altered body mass index: the authors speculate that palmitoylation stabilizes receptor localization at the cell surface. Further analysis is required to establish the consequences of MC₄ palmitoylation and identify the ZHHCs, as well as specific depalmitoylating enzymes, regulating this posttranslational modification.

Both MC₃ and MC₅ have 2 cysteines in their cytoplasmic tails, which are predicted to be palmitoylated (https://swisspalm.org/): at C315/C317 and C311/312, respectively. To date, there are no reports that describe these cysteines in any detail, however, mutation of the residue separating C315/C317 (pG316D) in MC₃, has been reported to result in a lean phenotype (6). We predict that this amino acid change is sufficient to prevent palmitoylation and hence anchoring of the cytoplasmic tail to the cell membrane, and that a similar mechanism may be essential for the normal function of several MCs. Diet, in particular fatty acids, has been shown to modulate palmitoylation (46), therefore, one might speculate that MC function may also be modified by diet.

Not All MC Signaling Is Mediated Through cAMP

Canonical MC Signaling

Initially, descriptions of MC activation concurred that all MCs are Gαs−coupled, activating adenylyl cyclase, which in turn catalyzes the conversion of ATP to cAMP. The second messenger cAMP will initiate an intracellular cascade, often through activation of protein kinase C (PKC). MC activation and signaling is terminated by recruitment of beta-arrestin, which traffics the receptor back to endosomes.

The affinities and potencies reported for the different ligands at each MC are highly variable between studies. To understand this variability, we systematically reviewed the literature reporting the cAMP response to different ligands for MC_1,3-5. We found 100-fold differences in the published EC50s (potencies) for MC_1,3-5 in response to the same ligand (Fig. 1). These fold differences were even found in data published by the same laboratories. As will be reviewed below, the cAMP response to different MC ligands is complex and context dependent. We suggest that the significant range of cAMP responses measured is a function of receptor interaction with other proteins and/or receptors.

Figure 1.

Reported potencies (EC50) of endogenous and exogenous melanocortin ligands for the melanocortin receptors in the published literature. To obtain the values, Web of Science was searched for “potenc*” or “ec50” and the names of the receptors using their various naming conventions. Only values obtained with the following methodologies were included: untagged receptor constructs transfected into a cell line and receptor activity measured in a cyclic AMP- or cyclic AMP response element (CRE)-based assay. Values from literature reviews were excluded. All values given for melanocortin receptor 1 (MC₁) are for the MC₁a isoform only. Abbreviations: ACTH, adrenocorticotrophic hormone; MC, melanocortin receptor; MSH, melanocyte-stimulating hormone; MTII, melanotan II; NDP, Nle4, D-Phe7.

Evidence for Signaling Through Other G Protein Alpha Subunits

Some of the MCs may interact with other Gα subunits: Gα _i and/or Gα _q/11 (MC₃ (47); MC₄ (48)). In neuronal cell culture, it has been demonstrated that activation of MC₄ by α-MSH in neurons of the paraventricular nucleus (PVN) results in activation of Gα _q/11 and not Gα _s (49). What has yet to be determined are the mechanisms that switch a MC from interacting with Gα _s to Gα _i or Gα _q/11.

Gα Independent Coupling With Kir7.1

Kir7.1 is an inwardly rectifying K⁺ channel and coupling with MC₁ and MC₄ has been demonstrated (50). In the PVN, the depolarization and hyperpolarization induced by α-MSH and AgRP, respectively, occurred independently of Gα pathways downstream of MC₄ (50). MC₄ appears to be unusual among the GPCRs assessed to date, in that it does not modulate the activity of Kir7.1 via glycosylation (51). Targeted deletion of Kir7.1 in MC₄-expressing cells of the PVN resulted in the failure of α-MSH to activate these MC₄ neurons and blocking of associated phenotypes (52). By contrast, the phenotypes associated with the activation of MC₄ by AgRP were not blocked. Recent tertiary structural analysis suggests that MC₄ signaling associated with coupling to Kir7.1 also requires Ca²⁺ binding (53). It is not yet known if coupling to Kir7.1 is a generic property of the MCs or unique to MC₄ and possibly MC₁.

Constitutive Activity of MCs

MCs appear to have constitutive cAMP-generating activity: the evidence for this being the case with MC₁ and MC₄ is the most compelling. Pomc KO mice maintain normal coat color even in the absence of endogenous ligands, while MC₁ knockout mice are yellow (phaeomelanin), suggesting that the MC₁ constitutive activity in the absence of endogenous ligands is sufficient to maintain coat color (54). MC₄ has some constitutive activity and AgRP is able to act as an inverse agonist in the presence of this activity (55).

Whether MC₃ is constitutively active is debatable: some report that the human MC₃ is not (56, 57) while others report some basal activity as measured by cAMP. The constitutive activity may therefore be species- and context-dependent (58). A mutant form of MC₃ (F347A) is constitutively active (58): its basal cAMP activity is about 7-fold greater than that that of wild-type MC₃. Even in the absence of ligand, mouse or human MC₅ stably transfected into B16/G4F melanoma or HEK293 cells, respectively, produced cAMP (40, 59), although this has not been detected by others (57).

In the late 1990s, it had been concluded that the N-terminus could be removed from all 4 receptors without effects on receptor function (60). However, later work on MC₄ showed that the N-terminus acts as a tethered ligand and is responsible for the constitutive activity of the receptor (55, 61). MC₄ constitutive activity has been shown to be augmented in the presence of human MRAPa (the long isoform of human MRAP1) and may be due to human MRAPa enhancing N-linked glycosylation of the N-terminus of MC₄ (57, 62). Constitutive activity may provide tone to a signaling pathway; that is, the ability to move in either direction from a set point.

Biased Signaling

The early understanding of the role of beta-arrestin in GPCR signaling was that it terminated the intracellular signaling cascade initiated by the Gα subunit. It is now understood that beta-arrestin can initiate its own signaling and this activity can occur once the activated receptor has been internalized and is in the early endosome.

Regardless of the usual Gα subunit that a GPCR normally activates, its associated beta-arrestin can also recruit Gα _i to form a complex that interacts with the ERK1/2 pathway (63). Others have previously demonstrated that MC activation can upregulate the MAPK ERK1/2 pathway independent of cAMP but dependent on PI3K (MC₁ (64, 65); MC₃ (66); MC₄ (67); MC₅ (68, 69)) and inhibit the MAPK c-Jun N-terminal kinase (JNK) pathways (MC₄ (70); MC₅ (71)).

Biased agonism by AgRP binding to MC₃ and MC₄ has been demonstrated (72, 73): activation of either receptor with AgRP independently stimulates the ERK1/2 pathway while decreasing cAMP activity. Whether any of these pathways are beta-arrestin dependent is unknown, as is the extent of biased signaling.

Modulators of MC Expression and Activation

There is accumulating evidence that several proteins can interact with the MCs to modulate their activation; including the MRAPs, membrane-bound attractin, mahogunin ring finger, and defensin. The latter 3 proteins will not be discussed in this review. MC activation may also be modulated by whether the receptor is acting as a monomer, homodimer, or heterodimer with other MCs or other GPCRs. What is not yet clear is how these putative modulators of MC activation exert their effects. It is possible that these modulators result in biased signaling and/or regulate the number of MCs presenting at the cell membrane.

MRAPs Modulate MC Activity

MRAP1 and MRAP2 interact with and regulate the function of all members of the MC family (74) as well as other GPCRs (75-77). While MRAP1 is present as antiparallel homodimers at the plasma membrane (78), MRAP2 can also form parallel homodimers as well as higher-order oligomers (79). Both MRAPs are widely expressed in several tissues, including the brain, pituitary, adrenal gland, testis, ovary, lung, and heart (74, 75, 80), which in part overlaps with expression of MCs. Coexpression of MRAP2 with MC₃ and MC₄ in the same cells has been demonstrated at the RNA level (81).

The complexity of the MC family is further evidenced by the contradictory influence of MRAPs on MC function, MC₃ being a prime example. Coexpression of human MRAP2 and MC₃ has been shown to either reduce (81) or have no influence (74) on MC₃ surface expression. Human MRAP1 and MRAP2 increased MC₃ cAMP signaling in response to α-MSH (57, 81), whereas human MRAP2 inhibited and MRAP1 did not significantly influence the MC₃ cAMP response to NDP-α-MSH in another study (74). Chicken MRAP2 produced a 9-fold increase of the potency of chicken ACTH (1-39) at MC₃ (82), whereas coexpression of chicken MRAP2 and MC₃ had no effect on the potency of human ACTH (1-24) (83). Zebrafish MRAP2a or MRAP2b had no significant effect on α-MSH-induced MC₃ cAMP signaling at a ratio of 1:5 of receptor to MRAP2 (84), whereas MRAP2 of the related channel catfish inhibited the cAMP response of MC₃ to α-MSH at the same receptor to MRAP2 ratio (85).

The divergent effects of the interaction between MRAPs and MCs highlight the influence of context on MC function. The concentration of MRAPs appears to be one of the context-dependent factors that influence MC activity, as MRAP2 alters receptor function differently depending on the expression ratio of MRAP2 to MC (81, 85, 86). What underlies the dose-dependent effects is unclear, however, the ability of MRAP2 to form different homo-oligomeric conformations, each with a potentially different effect on receptor function, may play a role (79, 86). As MRAP2 is differentially expressed across zebrafish development (84) and in the endometrium during different stages of the human menstrual cycle (87), altering the cellular concentration of MRAP2 may be an additional mechanism used by organisms to fine-tune MC signaling.

MC Homodimerization and Heterodimerization

All 5 MCs have the ability to homodimerize (88-91). While the exact cellular ratio of receptor monomers to homodimers is unclear, the prevalence of homodimers can be regulated by ligands and interacting proteins with potential functional consequences. ACTH binding increases MC₂ homodimerization (92), MRAP1 reduces the plasma membrane concentration of MC₅ by inhibiting MC₅ homodimerization (91) and disruption of MC₄ homodimers increases receptor-mediated cAMP accumulation (93). Furthermore, MC₄ has 2 tandem binding sites with different ligand binding affinities and kinetics, likely corresponding to sites on receptor homodimers (94). Homodimerization may therefore produce additional MC states with new functional properties and distinct interactions with other membrane proteins.

Bioluminescence resonance energy transfer and co-immunoprecipitation assays have provided evidence for physical association between different MCs. Heterodimerization between flounder MC₁ and MC₅ (95), human MC₁ and MC₃ (88) and mouse MC₃ and MC₄ (96) has been demonstrated in transfected cells. The receptor pairs are also coexpressed in vivo: MC₁ and MC₅ in flounder melanophores (97), MC₁ and MC₃ in alveolar macrophages (98) and MC₃ and MC₄ in the murine hypothalamus (81). Studies of the functional significance of MC interactions to date, indicate that any effects are highly ligand dependent. The efficacy of α-MSH in cells co-transfected with flounder MC₁ and MC₅ was significantly lower than in cells transfected with either MC₁ or MC₅, whereas the efficacy of desacetyl-α-MSH was significantly increased in double transfected compared to single transfected cells (95). Coexpression of mouse MC₃ and MC₄ had no significant effect on the potencies of α-MSH, NDP-α-MSH, or melanotan II, whereas the potency of bivalent ligand CJL-1-87 was moderately increased in cells expressing both MC₃ and MC₄ compared to a mixture of cells expressing MC₃ and MC₄ separately (96).

The Interaction Partners of MCs Are Not Limited to Members of the MC Family

Recently, Li et al provided a significant advance toward characterizing the protein interactomes of MC₃ and MC₄ by identifying 23 and 32 GPCRs, respectively, that physically associate with the 2 MCs in vitro (99). The functional consequences of receptor coexpression were diverse, with inhibition, potentiation, and no effect on MC₃ and MC₄ signaling observed depending on the GPCR partner present. Previous studies have also described interactions between various GPCRs and MC₃ and MC₄ (100, 101). Despite the attested ability of the receptors to heterodimerize at the membrane, the reported effects of receptor coexpression on signaling activity may also arise due to crosstalk between signaling pathways. Such a mechanism may account for the combined effects of α-MSH and endothelin-1 on melanocyte function (102) and the signaling crosstalk between MC₃ and the growth hormone (GH) secretagogue receptor (103).

Future Perspectives

Understanding MC Signaling

As described above, there are many examples of noncanonical MC signaling and therefore understanding what factors determine MC signaling is critical for optimizing the selectivity and efficacy of pharmacological interventions. Recently, setmelanotide received FDA approval for chronic weight management for patients with, in effect, genetic ablation of POMC, PCSK1 (proprotein convertase subtilisin/kexin type 1, responsible for cleavage of POMC resulting in α-MSH and ACTH), or LEPR (leptin receptor). A series of clinical studies demonstrated significant weight loss in these patients and chronic treatment was not associated with the negative side effects seen with the use of other agonists (104-106). Its use has not been without off-target effects though. Individuals with genetic ablation of POMC and PCSK1 are characteristically fair with red hair. After extended treatment with setmelanotide, the hair color of these individuals became brown, demonstrating that the setmelanotide is also acting on MC₁ (107). This is not surprising since setmelanotide is known to also interact with both MC₁ and MC₃ albeit with lower potencies (108). Setmelanotide has biased Gα _q/11 signaling and binds to MC₄ differently to α-MSH (36). Although the identification of a highly selective agonist or antagonist for any of the MCs remains elusive, compounds like setmelanotide provide insight into the structure and function of MCs, in addition to having therapeutic use.

Understanding GPCR Crosstalk

To date, research on GPCR crosstalk has mostly been limited to interactions between 2 partners due to the lack of techniques for detecting large multi-member protein oligomers (109) and the challenge of untangling the complex functional effects caused by interplay between several proteins. New techniques have generated demonstrations of the formation of higher-order receptor oligomers (109-112). Such “receptor mosaics” (113) may not only be composed of several different GPCRs but also of accessory proteins, ion channels and other types of receptors which together determine the functional properties of the larger unit (114).

The diverse interaction profiles of the MCs suggest that the receptors participate in larger heteromeric complexes. The GPCRs interacting with MC₃ and MC₄ are all expressed in the hypothalamus, many of them in the same cells (99). Given that several of the GPCRs heterodimerize with each other in addition to interacting with MCs, the number of possible oligomeric complexes that may form is staggering. Which of these interactions occur in vivo and what determines the oligomeric species present at any one time remains unclear, however, the cellular context is likely to have a major influence. Complex interactions between MCs and the many different proteins that make up the cellular environment may therefore give rise to context-dependent functional units, which each respond to ligands in a unique manner. Differential expression of some of these functionally diverse MC complexes between different cell lines and cell states may help explain the variable potencies reported for MCs.

What Is the Role of Nanodomains in Producing Context-Dependent Functional Units?

Within a cell, there is the possibility of a variety of nanodomains (115): a localized membrane environment that may contain receptor mosaics, hetero- or homodimers of the MCs along with different G proteins, beta-arrestins, and accessory proteins.

MC₃ transfected into a mouse neuronal cell line localizes to lipid rafts (116), one type of nanodomain. Organization of different MC oligomers into distinct nanodomains could provide spatial separation of signaling responses and contribute to the diverse MC responses observed, since the distribution and makeup of lipid rafts is heterogeneous between and within cell types (117, 118). The importance of MC compartmentalization has already been shown for MC₄, which requires MRAP2-mediated trafficking into primary cilia in the PVN for its anorexigenic effect (119, 120). The presence of MC₄ on primary cilia may also be indicative of another role; that is, that the MCs may be involved in volume transmission (121).

Understanding how these different membrane proteins are compartmentalized to different nanodomains, and the potential role for posttranslational modifications, such as palmitoylation, is going to be essential to understanding the diversity of responses following MC activation.

Do Specific MCs Have Roles in a Broader Range of Tissues?

In order to advance our understanding of MC biology, further research into their exact cellular localization throughout the body is required. The development of a “rainbow” mouse expressing each of the MCs tagged with a different fluorophore might therefore be useful. Determining the exact cellular localization of each of the potential interactors with the MCs is also required and whether physiological status changes the combinations of interactors. In the absence of specific antibodies for each of the MCs, this may be possible by multiplex RNAscope combined with tissue optical methods (for example, see (122)). The interactions of MCs with other GPCRs suggests they may act as the conductor of an orchestra by monitoring the activity of these other GPCRs, perhaps through oligomerization, to regulate their signaling and therefore cellular responses. Further structural studies to determine the tertiary structures of each of the MCs with and without either MRAP, as well as with and without different receptor dimers, are also required. Countless facets of MCs remain to be fully explored and understood.

Conclusions

The focus of a wide body of research on specific roles for the MCs attests to their importance in physiology, however, their potential importance in the function of a range of other tissues is currently unclear. This has implications both for understanding mechanisms leading to disease as well as for characterization as therapeutic targets. Study of these other roles, as well as those that are well established, will require a more complete understanding of their multifaceted biology and how this relates to ligand specificity, as well as modulation of signaling from these GPCRs in the context of their interacting partners.

Glossary

Abbreviations

α-MSH

α-melanocyte stimulating hormone

ACTH

adrenocorticotropic hormone

AgRP

agouti-related peptide

cAMP

cyclic adenosine monophosphate

GPCR

G protein–coupled receptor

KO

knockout

MAPK

mitogen-activated protein kinase

MC

melanocortin receptor

MRAP

melanocortin receptor accessory protein

POMC

proopiomelanocortin

PCSK1

proprotein convertase subtilisin/kexin type 1

PVN

paraventricular nucleus

TMD

transmembrane domain

Funding

Project support grant from the British Society of Neuroendocrinology (LL).

Disclosures

The authors have no conflicts to disclose.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.