Agouti: from Mouse to Man, from Skin to Fat


  • Joanne Voisey,

    1. Co-operative Research Centre for Diagnostics, Queensland University of Technology, Brisbane, Australia
    Search for more papers by this author
  • Angela Van Daal

    1. Co-operative Research Centre for Diagnostics, Queensland University of Technology, Brisbane, Australia
    Search for more papers by this author
    • *Address reprint requests to Dr. Angela van Daal, Co-operative Research Centre for Diagnostics, School of Life Sciences, Queensland University of Technology, 2 George Street, Brisbane 4000, Australia. E-mail:


The agouti protein regulates pigmentation in the mouse hair follicle producing a black hair with a subapical yellow band. Its effect on pigmentation is achieved by antagonizing the binding of α-melanocyte stimulating hormone (α-MSH) to melanocortin 1 receptor (Mc1r), switching melanin synthesis from eumelanin (black/brown) to phaeomelanin (red/yellow). Dominant mutations in the non-coding region of mouse agouti cause yellow coat colour and ectopic expression also results in obesity, type II diabetes, increased somatic growth and tumourigenesis. At least some of these pleiotropic effects can be explained by antagonism of other members of the melanocortin receptor family by agouti protein. The yellow coat colour is the result of agouti chronically antagonizing the binding of α-MSH to Mc1r and the obese phenotype results from agouti protein antagonizing the binding of α-MSH to Mc3r and/or Mc4r. Despite the existence of a highly homologous agouti protein in humans, agouti signal protein (ASIP), its role has yet to be defined. However it is known that human ASIP is expressed at highest levels in adipose tissue where it may antagonize one of the melanocortin receptors. The conserved nature of the agouti protein combined with the diverse phenotypic effects of agouti mutations in mouse and the different expression patterns of human and mouse agouti, suggest ASIP may play a role in human energy homeostasis and possibly human pigmentation.


untranslated region


base pairs


long interspersed elements


epidermal growth factors


central nervous system


Two key genes, agouti and extension, are involved in melanin switching in the mouse. The extension gene encodes the melanocortin 1 receptor (Mc1r) while the agouti gene encodes a 131 amino acid protein consisting of a signal peptide, glycosylation site, a basic domain and a cysteine rich C terminus (1). The agouti gene has been mapped and isolated in a number of species. It has been demonstrated to play a role in coat colour determination in sheep, standard silver fox and rat and has also been isolated in pig (2–7). The human homologue of agouti is located on chromosome 20 and encodes a 132 amino acid protein, agouti signal protein (ASIP) (8, 9). Despite the high degree of protein homology (80% of amino acids identical to mouse) the expression pattern is quite different. Agouti is normally expressed in the hair follicles in neonatal mice and no agouti expression is found in a variety of mouse adult tissues including brain, liver, lung, spleen and kidney although it is expressed in adult testis (1). In humans, ASIP is also expressed in testis as well as ovary, heart, adipose tissue and at lower levels in liver, kidney and foreskin (8, 9).

This review is concerned primarily with the role of ASIP in humans and how expression affects pigmentation and obesity. Although we have some understanding of the role that agouti plays in the mouse, a role for ASIP in humans is yet to be determined. Throughout this review, the mouse gene will be referred to as agouti whereas its human orthologue will be referred to as ASIP. The protein will be referred to as agouti in mouse and ASIP in humans.


Regulation of Pigmentation Markers by Agouti

In mice signalling by α-melanocyte stimulating hormone (α-MSH) or agouti through Mc1r regulates the type of melanin produced. α-MSH interaction with Mc1r results in eumelanin production and agouti antagonism of this interaction results in phaeomelanin production. Three novel genes have been found to be up-regulated by agouti protein, including a potential DNA replication control protein (minichromosome maintenance protein, MCM6), a basic helix-loop-helix transcription factor (immunoglobulin transcription factor 2, ITF2) and a gene of unknown function expressed in retinal cells (10). Agouti protein also down-regulates genes to inhibit eumelanin synthesis. Treatment of mouse melanocytes with agouti protein results in decreased expression of genes such as, tyrosinase related protein 1 (Tyrp1) and dopachrome tautomerase (Dct) and to a lesser extent, tyrosinase (11). Following exposure to α-MSH, tyrosinase expression increased in melanocytes but no effect was evident for Tyrp1 or Dct. Similar effects are seen with mouse melanoma cell lines (12).

The role of ASIP in pigmentation in humans has not been extensively investigated. In vitro studies have shown that human ASIP acts in a similar fashion to mouse agouti protein in that it down-regulates melanin genes controlling eumelanin synthesis in human melanocytes (13). This suggests a role for ASIP in human pigmentation, however, further studies are required to determine which cell type in the skin expresses ASIP and if ASIP alters the expression of other pigmentation genes. Characterizing the biochemical mechanisms that cause genes to switch to phaeomelanin production is important in photo-protection and the identification of risk factors.

Agouti and MC1R

Melanocortin 1 receptor has been the only gene associated with normal hair colour variation in humans to date. Variants of the MC1R gene (R151C, R160W, D294H, R142H, 86insA and 537insC) are associated with lighter skin types and red hair (14–18). The involvement of another locus in the inheritance of red hair is evidenced by the existence of redheads carrying no MC1R variants (14, 17) and some dizygotic twin pairs are concordant for MC1R variation, but discordant for hair colour (16). Human variation in hair and skin pigmentation may result from a combination of ASIP and MC1R alleles, as is the case for the German shepherd dog (19), with some MC1R variants interacting differently with ASIP resulting in different ratios of eumelanin and pheomelanin production.

Agouti Mutations

Five dominant mutations result in ectopic expression of the agouti protein in mice. These produce pleiotropic effects, including yellow coat colour, obesity, diabetes and tumour susceptibility. The lethal yellow (Ay) mutation, but not the other dominant agouti mutations, results in embryonic lethality in the homozygous state (1, 20, 21). The Ay genotype results from a 120–170-kb deletion that removes the entire coding region of the upstream raly (heterogeneous nuclear ribonucleo-protein associated with lethal yellow) gene as well as the non-coding exon of agouti. As a result, the Ay allele is under the control of the raly promoter, which is normally ubiquitously expressed (22–24). Like Ay, the other dominant agouti mutations, intracisternal A-particle yellow (Aiapy), intermediate yellow (Aiy), sienna yellow (Asy), and viable yellow (Avy) result in ubiquitous expression, but these contain insertions of varying sizes upstream of the first coding exon (25–27). Transgenic mice that ectopically express agouti protein have yellow fur, are obese and develop insulin resistant diabetes (28) suggesting that the pleiotropic effects associated with dominant yellow mutations are caused by ectopic agouti expression.

Recessive agouti mutations in the mouse are associated with increased eumelanin and decreased phaeomelanin production and include non-agouti (a), black and tan (at), lethal non-agouti (ax), non-agouti lethal (al), tanoid (atd) and extreme non-agouti (ae) (29–31). To date, no ASIP polymorphisms have been identified in humans. A polymorphism study of defined ethnic groups, varying in skin and hair pigmentation, revealed no polymorphisms in the ASIP coding region and also showed no polymorphisms in individuals with MC1R genotypes inconsistent with phenotypes (32). This leaves open the possibility that polymorphisms associated with human pigmentation exist in the non-coding region of ASIP.

Agouti Expression

Agouti is the murine wild-type coat colour, characterized by a black hair with a subapical yellow tip. Transient expression of agouti results in the switching of eumelanin to phaeomelanin from days 4–6 of the hair cycle and back to eumelanin after day 6. The switch to production of phaeomelanin is regulated by agouti antagonism of α-MSH signalling through Mc1r (33). Immunohistochemistry studies of mouse skin with agouti revealed high levels of protein expression in 6-day wild-type agouti mice but not in non-agouti mice. In heterozygous lethal yellow mice, the level of expression correlated with coat colour intensity, indicating that the agouti protein is localized within the follicular melanocyte, where it stimulates phaeomelanin production and inhibits eumelanin production (34, 35). Transgenic mice expressing agouti in the basal epidermal cells develop bands of yellow hair which correspond to the regions of epidermal agouti expression, suggesting agouti protein functions in a paracrine manner in the mouse (36).

Agouti phenotypes result from a single coding sequence regulated by alternative promoters that control coat colour of dorsum and ventrum independently (29, 37) (Fig. 1). The hair specific promoter is activated from days 4–6 of the hair cycle in both dorsal and ventral skin and results in the characteristic subapical yellow tip. The ventral-specific promoter is activated throughout the hair cycle producing yellow hair but only in the ventrum. Mutations affecting the hair specific promoter such as black and tan (at) result in mice with a yellow ventrum and black dorsum.

Figure 1.

 Alternative agouti promoters. Vrieling et al. (37) isolated cDNA from total skin RNA of 5-day-old mice. The majority of cDNA clones contained three coding exons and one non-translated 5′ exon. However, a few clones had a different 5′ end. To investigate this, cDNA was isolated from ventral skin RNA of 1-day-old light-bellied agouti mice. Four RNA isoforms were present. Exon 1A and 1A′ are located ˜120 kb upstream of exon 2 and are present only in agouti mRNA from the ventral skin. Exons 1B and 1C (hair specific) are ˜16 nucleotides apart and are located ˜18 kb upstream of the exon 2.


The melanocortin family consists of five genes encoding G protein-coupled receptors, MC1R–MC5R which all have seven transmembrane domains. In humans MC1R is expressed in melanocytes, MC2R is expressed in adrenal and adipose tissue, MC3R is expressed in brain, placenta and pancreas, MC4R is expressed in brain, muscle and adipose tissue and MC5R is expressed in most tissues (38). The expression pattern for mouse melanocortin receptors is similar (39). A key difference is that MC4R in the mouse is expressed in the brain but not in adipose tissue (40) whereas in humans, expression occurs in both adipose tissue and brain (41). Together, the melanocortin family of receptors regulate pigmentation, lipolysis, food intake, thermogenesis, sexual behaviour, memory, anti-inflammatory and antipyretic effects (42). α-MSH acts as a ligand for MC1R, MC3R, MC4R and MC5R; γ-MSH is a ligand for MC3R and adrenocorticotropic hormone (ACTH) is a ligand for MC1R, MC2R, MC4R and MC5R (Fig. 2). ACTH, α-MSH and γ-MSH are all derived from the same precursor protein, preproopiomelanocortin (POMC). Mouse agouti protein antagonizes Mc1r and Mc4r and to a lesser extent Mc3r but has no effect on Mc5r (33, 43). Because of difficulties in generating heterologous cell lines that express significant levels of Mc2r, antagonism of Mc2r by agouti has not previously been reported. However, using an adrenocortical cell line (OS3), Yang et al. (44) has shown that human ASIP is an inhibitor of human MC2R. Human ASIP is also a potent inhibitor of human MC1R, MC4R and has weak effects at MC3R and MC5R (44). Differences in the patterns of mouse and human antagonism may be explained by differences between sequences or more likely, differences in the concentrations of recombinant agouti protein used in each study. Results by Yang et al. (44) indicate ASIP antagonizes all the known melanocortins with varying affinities, suggesting an even wider function for ASIP in humans. Studies of agouti antagonism of melanocortin receptors provide a model for identifying a function for ASIP in humans. The ability of ASIP to antagonize MC1R, MC3R and MC4R as well as expression of ASIP in skin and adipose tissue in humans suggests a possible key role for ASIP in obesity, diabetes and pigmentation with each receptor interaction having an important physiological role.

Figure 2.

 Agonism and antagonism of the melanocortin receptors. Adrenocorticotropic hormone (ACTH), α-melanocyte stimulating hormone (α-MSH) and γ-MSH are ligands of the melanocortin receptors; human agouti signal protein (ASIP) antagonizes all five melanocortin receptors to varying degrees. Human ASIP is a potent antagonist of MC1R and MC4R and a weak antagonist of MC2R, MC3R and MC5R as indicated by the dashed lines. Together, the melanocortin receptors are widely expressed and regulate pigmentation, lipolysis, food intake, thermogenesis, sexual behaviour, memory, anti-inflammatory and anti-pyretic effects.

The carboxyl terminus residues 83–131 of agouti are as potent in antagonizing melanocortin receptors as the full-length protein (45) and alanine scanning mutagenesis indicates that amino acid Arg116, Phe117 and Phe118 are critical to the affinity of agouti for Mc1r, Mc3r and Mc4r (46). Not surprisingly these three residues have chemical properties that resemble the core sequence (His-Phe-Arg-Trp) of the melanocortin peptides (46). In the mouse, binding of α-MSH to Mc1r is antagonized by agouti, resulting in decreased cAMP levels and a switch from eumelanin to phaeomelanin synthesis. The question of whether agouti protein binds directly to Mc1r in mammalian melanocytes or acts on another receptor has been addressed. Responsiveness to agouti protein in melanocytes from mice expressing a functional or mutant Mc1r gene showed that a functional Mc1r is required for agouti to exert its effects as measured by decreased levels of tyrosinase, Tyrp1 and Dct (47). The antagonistic effects of agouti protein on α-MSH activity have been studied in vivo by Ollmann et al. who demonstrated that a functional Mc1r receptor is required for agouti signalling in vivo and that the effects of agouti are not explained solely by inhibition of α-MSH (48). Studies using mouse melanoma and human melanocyte models have shown that human ASIP can act independently of α-MSH to switch melanin synthesis from eumelanin to phaeomelanin (49). In the absence of α-MSH, ASIP inhibits melanogenesis but has little effect on the ability of α-MSH to induce melanogenesis, suggesting phaeomelanin production may not solely be the consequence of ASIP antagonizing MC1R. A second receptor may therefore be involved in agouti signalling. The mouse mahogany locus, which suppresses the coat colour and obesity phenotypes of the lethal yellow mouse, has recently been shown to encode the membrane form of attractin and could be another receptor for agouti signalling.

Agouti and Attractin

Human attractin was first identified as a serum glycoprotein that is expressed on activated T cells, resulting in an association between T cells and monocytes (50). It is subsequently released to the extracellular environment and mediates the spreading of monocytes and clustering of T cells. Genomic sequencing revealed that the soluble form of attractin is not a proteolytic product of the membrane form, but that both the secreted and membrane forms of attractin result from alternative splicing of the same gene (51). The soluble attractin form results from transcription of 25 sequential exons with the last exon containing a LINE-1 retrotransposon element with a stop codon and polyadenylation signal. The membrane form of attractin is generated by splicing over this exon and transcription of a further five exons that encode the transmembrane and cytoplasmic domains. The attractin proteins contain dipeptidyl peptidase activity, four EGF-like motifs which are characteristic of proteins involved in extracellular signalling or cell guidance, a CUB domain and a C type lectin domain, both of which are characteristic of proteins that interact with carbohydrate moieties as well as a domain homologous with the ligand-binding region of the common γ cytokine chain. Both forms of attractin are widely expressed suggesting a wide spectrum of physiological functions.

The demonstration that the mouse mahogany and the rat zitter loci both encode attractin has led to speculation about the role of this protein and its interaction with agouti. The mouse mahogany mutation, recently renamed Atrnmg, suppresses the yellow coat colour and obesity of Ay mice suggesting it is a downstream mediator of agouti signalling (52–54). Atrn is believed to act upstream of the melanocortin receptors as it fails to reverse the respective pigmentation and obesity effects of Mc1r and Mc4r knockout mice (53–55). Mice homozygous for Atrnmg and the loss-of-function Mc1re allele are phenotypically indistinguishable from homozygous Mc1re mice confirming that Atrn is genetically upstream of Mc1r. While Atrnmg suppresses the obesity phenotype of Mc4r deficient mice, it does not suppress obesity caused by mutations in the leptin receptor (Leprdb), carboxypeptidase E (cpefat) or tubby indicating that the Atrn protein functions with some specificity in the melanocortin signalling pathway.


Mouse Model

One of the pleiotropic effects of the mouse agouti mutations is obesity with the degree of weight increase being proportional to the intensity of the yellow coat (56). Transgenic mice expressing agouti under the control of a ubiquitous promoter become obese, develop diabetes, tumours and a yellow coat (28). It is likely that Mc4r is responding to ectopic agouti protein expression, resulting in weight gain. Agouti expression in skin has little effect on the obesity phenotype and agouti protein does not circulate as an endocrine factor as transgenic mice over-expressing agouti in skin do not become obese (57). Agouti is not the normal antagonist of Mc4r but novel expression in the brain may result in the antagonism of Mc4r. Mc4r is expressed in the hypothalamus and plays a role in the regulation of feeding and metabolism (58). Mc4r knockout mice are obese but do not have a yellow coat. It seems likely that ubiquitous expression of agouti protein results in chronic antagonism of Mc4r thereby disrupting its function. Agouti antagonism of Mc4r is unlikely however, to be solely responsible for the obesity phenotype as administration of a melanocortin agonist to mice over-expressing agouti protein does not reverse the obesity phenotype (59). Administration of an Mc4r agonist does however, inhibit feeding and administration of an Mc4r antagonist increases feeding suggesting antagonism of Mc4r does play some role in agouti-induced obesity (60).

More recently a role for Mc3r has been implicated in body weight regulation. Mc3r knock-out mice have increased fat mass, reduced lean mass and a greater feeding efficiency compared with wild-type mice (61). In addition, knockout mice for both Mc3r and Mc4r gain significantly more weight compared with Mc4r knockout mice. It is believed that Mc4r knockout mice have increased feeding and Mc3r and Mc4r knockout mice are more efficient at storing calories.

Agouti-Related Protein (Agrp)

An Agouti-Related Protein (Agrp) cDNA has been identified through sequence homology to agouti (62, 63). Like agouti, Agrp encodes a melanocortin receptor antagonist that antagonizes Mc3r and Mc4r (63). It is likely that Agrp is the natural antagonist of Mc3r and Mc4r as it is expressed in the hypothalamus. Transgenic mice over-expressing Agrp develop obesity and have similar phenotypes to agouti transgenics (63) but do not develop a yellow coat as Agrp does not antagonize Mc1r. In addition, administration of Agrp stimulates feeding, mimicking the effects of Mc3r and Mc4r antagonists (64).

Human AGRP and ASIP have the same intron/exon structure, are the same length and contain signal peptides and cysteine rich carboxyl regions. However the basic domain found in ASIP is not present in AGRP and there is low sequence homology, except in the carboxyl terminus that is essential for agouti antagonist binding, where nine out of 11 cysteine residues are conserved (62, 63). AGRP may offer potential therapy in the treatment of human obesity. Because MC4R is widely expressed in the brain it is likely to be a regulator of other important functions. AGRP, however, is expressed primarily in the hypothalamus and so administration of an appropriate AGRP inhibitor may affect energy homeostasis without altering other physiological functions (65). AGRP has also recently been implicated in susceptibility to anorexia nervosa. Anorexic patients had a significantly increased frequency of an Ala67Thr polymorphism in the AGRP gene (11% compared with 4.5%; P=0.015) (66). It is likely that the polymorphism in AGRP disrupts MC4R signalling, causing a decrease in feeding. Antagonism of MC4R by another functional ligand such as ASIP should counter these effects to increase feeding and may therefore provide a basis for treatment of anorexia.

Regulation of Adipocity by Agouti

The hormone leptin is produced in adipose tissue and circulates to the brain where it binds to the leptin receptor, signalling a decrease in food intake and an increase in energy metabolism thereby limiting the degree of obesity (67, 68) (Fig. 3). Leptin levels are increased in adipose tissue in agouti transgenic mice, suggesting that agouti may interact with other obesity gene products to play a key role in weight regulation (69). Agouti has indeed been demonstrated to be involved in the regulation of leptin in the lethal yellow mouse (69).

Figure 3.

 Antagonism of MC4R in humans. α-Melanocyte stimulating hormone (α-MSH) binds to MC4R in the hypothalamus, regulating energy homeostasis. AGRP, also expressed in the hypothalamus, antagonizes MC4R, altering weight regulation. Agouti signal protein (ASIP) normally expressed in adipose tissue in humans may also antagonize MC4R effecting energy homeostasis.

The body weight of transgenic mice expressing high levels of agouti in adipose tissue and low levels in other tissues is unaltered (70). However, when given daily insulin injections a significant increase in weight gain was seen, suggesting that the insulin-induced weight gain is mediated via agouti. It is likely that ectopic expression of agouti in the pancreas stimulates insulin release in the obese yellow mouse. It is worth noting that humans normally express ASIP in adipose tissue (8, 9) and in lower levels in the pancreas (71) indicating that ASIP and insulin together may regulate adiposity in humans.

Lethal yellow mice also have increased levels of fatty acid synthetase (FAS) and stearoyl-CoA desaturase (SCD), which are involved in fatty acid synthesis and desaturation of saturated fatty acids, respectively, relative to lean controls (72). FAS and SCD mRNA levels increased 1.5- and 4-fold, respectively, in 3T3-L1 adipocytes treated with recombinant agouti protein, suggesting that over-expression of FAS and SCD in lethal yellow mice is regulated by increased leptin induced by agouti (72). Lethal yellow mice have increased adipocyte size, indicating that increased lipid synthesis in adipocytes may contribute to the obese phenotype. As humans normally express ASIP in adipose tissue, ASIP may have a direct or indirect effect on adipocyte size in humans. In addition, adipose tissue ASIP mRNA levels have been positively correlated with FAS levels in humans (73). In 3T3-L1 adipocytes, agouti protein and insulin independently and additively increased FAS expression and in human adipocytes insulin and agouti are both required for FAS elevation (74). In addition, transgenic mice, over-expressing agouti in adipose tissue, fail to gain weight unless insulin is administered (70). These studies confirm that insulin and agouti protein regulate lipogenesis in an additive manner to increase body weight. More recently agouti has been shown to up-regulate adipocyte transcription factors, signal transductors and activators of transcription (STAT 1 and 3) and peroxisome proliferator-activated receptor (PPAR)-γ protein in 3T3-L1 adipocytes and in transgenic mice expressing agouti in adipose tissue (60).

Agouti protein stimulates lipogenesis in adipose tissue via calcium signalling (71). Recombinant agouti protein causes dose-dependent increases in intracellular calcium concentrations in mouse and human adipocytes, signalling the release of insulin (75, 76). In addition, in pancreatic β-cell lines as well as human pancreatic islets, agouti protein increased calcium levels and insulin release, suggesting ASIP may function in insulin regulation in humans (77).

Agouti and Attractin in Obesity

Mutations at the mouse Atrn or the rat zitter locus result in a phenotype whereby animals show neurological defects manifested as body tremors and progressive hypomyelination and vacuolation in the central nervous system (CNS). It is known that Atrn controls feeding and metabolism independently of agouti suppression (52). Atrn mutant mice eat more than normal mice but do not gain weight, possibly because they have increased metabolism and motor activity (52, 54). Interestingly the only abnormal effect seen by He et al. outside the CNS in Atrnmg-3J homozygous mice was reduced fat storage in both brown and white adipose tissue (78).

Recent studies have shown that Atrn is a receptor for agouti but not Agrp. Interaction of Atrn occurs through the N terminal region of agouti, which has low sequence homology to Agrp whereas the melanocortin receptor interaction occurs through the C-terminal region which has significant sequence homology to Agrp. In vitro and in vivo experiments show loss of Atrn suppresses agouti induced obesity but not Agrp-induced obesity (78). These results suggest that Atrn binding to the N-terminal region of agouti is required for agouti protein signalling mediated through Mc1r. Atrn and agouti binding may also be critical for Mc4r signalling in the brain.

While in the mouse agouti is not normally expressed in adipose tissue and is therefore unlikely to play a role in normal energy homeostasis, ASIP is expressed in human adipose tissue. A role for ASIP in energy homeostasis is implicated in signalling through attractin and the melanocortin signalling pathway. Further studies are needed to determine the exact interactions between Atrn and melanocortin receptors in humans.

Human Model

A possible role of ASIP in humans is the regulation of energy homeostasis by antagonism of MC4R. As discussed above, Mc4r plays a key role in the regulation of feeding as Mc4r knock-out mice are obese compared with controls and heterozygous Mc4r knock-out mice are intermediate in body weight with females being more susceptible to weight gain (79). Loss-of-function mutations of MC4R have also been identified in obese humans (80, 81). In a cohort of 63 obese children [mean body mass index (BMI), 34 kg/m2], one subject was heterozygous for a 4-bp deletion, resulting in a non-functional receptor. This same mutation was inherited in a dominant manner from the obese father (BMI, 41 kg/m2).

Human obesity is likely to result from rare mutations in a number of different genes. MC4R appears to be one such gene as it shows a high prevalence of mutations in obese individuals. In a large French population of obese individuals (n=209), 4% carried rare heterozygous MC4R mutations, whereas these same mutations were not detected in any of the non-obese controls (n=254) (82). Polymorphisms of MC4R are the most frequent demonstrated cause of genetic obesity to date.

Mutations in Pre-ProOpiomelanocortin (POMC), the precursor protein of α-MSH have also been associated with obesity. Two unrelated individuals found to have defects of the POMC gene, exhibited adrenal insufficiency, red hair and early onset obesity (83). Adrenal insufficiency is probably caused by a defect in MC2R signalling, red hair the result of the inhibition of α-MSH binding to MC1R and obesity caused by defects in MC4R signalling. In addition, mice lacking POMC-derived peptides (ACTH, α-, β-, and γ-MSH and β-endorphin) have phenotypes similar to human POMC mutants including obesity, adrenal insufficiency and a yellow coat (84). Treatment with a α-MSH agonist decreased obesity in these mice, indicating a possible therapeutic use of melanocortins in the treatment of obesity. It is unlikely that mutations of POMC are a common cause of obesity in humans as mutation screening of 96 obese patients revealed no polymorphisms associated with obesity (85).

The yellow obese model in the mouse is particularly useful for studying human obesity, partly because expression in a novel tissue (such as adipose or the hypothalamus) leads to antagonism of Mc4r. To date, there have been no reported polymorphism studies of ASIP and its association with obesity in humans. Mutations in humans causing ubiquitous ASIP expression may result in chronic antagonism of MC4R in the hypothalamus disrupting its function in weight regulation (Fig. 3). It is also possible that wild-type ASIP may have a direct role in fat metabolism in humans as both ASIP and MC4R are expressed in adipose tissue.


Given the range of agouti phenotypes identified in the mouse it is likely that ASIP plays one or more significant roles in humans. Although the structure and sequence of human ASIP is homologous to the mouse, there is considerable variation in the tissues in which it is expressed. In humans, ASIP is expressed at highest levels in adipose tissue, while MC4R is expressed in the brain but also at lower levels in adipose tissue. Whether ASIP antagonizes MC4R in adipose tissue or acts on another receptor is a question remaining to be answered. Attractin has also been demonstrated to have a role in energy homeostasis as well as being a receptor for agouti. It therefore seems probable that ASIP plays some role in human energy homeostasis. As ASIP has been shown to be an antagonist of all the human melanocortins it may potentially affect a range of physiological functions such as motor skills, sexual behaviour, memory, thermogenesis, lipolysis as well as pigmentation and food intake. Sexual behaviour, for example, may be regulated by ASIP via antagonism of MC2R, which is expressed in the adrenal glands. ASIP may also antagonize MC2R in the kidneys where it is expressed at low levels to regulate steroidogenesis. It is likely that expression studies and biochemical studies will help elucidate the physiological roles of ASIP in the range of human tissues, including adipose tissue, ovary and testis, in which it is expressed.


  1. *Address reprint requests to Dr. Angela van Daal, Co-operative Research Centre for Diagnostics, School of Life Sciences, Queensland University of Technology, 2 George Street, Brisbane 4000, Australia. E-mail:


This work was supported by a Cooperative Research Centre for Diagnostic Technologies Award to JV. The authors would like to thank Associate Professor Phillip Morris, Dr. Flavia Huygens and Dr. Adele Millis for critical review of the manuscript.