Lynn E. Pritchard

ABSTRACT Proopiomelanocortin (POMC) can be processed to ACTH and melanocortin peptides. However, processing is incomplete in some tissues, leading to POMC precursor release from cells. This study examined POMC processing in human skin and the effect of POMC on the melanocortin‐1 receptor (MC‐1R) and melanocyte regulation. POMC was secreted by both human epidermal keratinocytes (from 5 healthy donors) and matched epidermal melano‐cytes in culture. Much lower levels of α‐MSH were secreted and only by the keratinocytes. Neither cell type released ACTH. Cell extracts contained significantly more ACTH than POMC, and α‐MSH was detected only in keratinocytes. Nevertheless, the POMC processing components, prohormone conver‐tases 1, 2 and regulatory protein 7B2, were detected in melanocytes and keratinocytes. In contrast, hair follicle melanocytes secreted both POMC and α‐MSH, and this was enhanced in response to corti‐cotrophin‐releasing hormone (CRH) acting primarily through the CRH receptor 1. In cells stably trans‐fected with the MC‐1R, POMC stimulated cAMP, albeit with a lower potency than ACTH, α‐MSH, and β‐MSH. POMC also increased melanogenesis and dendricity in human pigment cells. This release of POMC from skin cells and its functional activity at the MC‐1R highlight the importance of POMC processing as a key regulatory event in the skin.—Rousseau, K., Kauser, S., Pritchard, L. E., Warhurst, A., Oliver, R. L., Slominski, A., Wei, E. T., Thody, A. J., Tobin, D. J., White, A. Proopiomelanocortin (POMC), the ACTH/melanocortin precursor, is secreted by human epidermal keratinocytes and mela‐nocytes and stimulates melanogenesis. FASEB J. 21, 1844–1856 (2007)

Proopiomelanocortin (POMC) is processed in an intracellular secretory pathway, primarily to enable release of ACTH from the pituitary and alpha-MSH from hypothalamic neurons and skin. However, processing is incomplete and unprocessed POMC is secreted from all three tissues. This review considers intracellular processing of neuronal POMC as a key checkpoint that controls flux through hypothalamic melanocortin receptor pathways. Regulation of the convertase, proprotein convertase (PC)-1/3, which cleaves POMC is likely to determine the extent of POMC processing. Reduced PC1/3 activity, in both humans and rodents, leads to reduced melanocortin signaling and hence obesity. In contrast to POMC, posttranslational processing of proagouti-related peptide, an endogenous melanocortin-4 receptor antagonist, is efficient and is unlikely to represent a regulatory checkpoint. Because POMC is fully processed to ACTH and MSH peptides in secretory vesicles, unprocessed POMC, which is released from cells, must exit via an unregulated constitutive pathway. Therefore, the targeting of POMC to secretory granules controls the extent of POMC cleavage. There is evidence that PC1/3 is involved in cleavage of POMC in the trans-Golgi network and regulation of trafficking to the secretory pathway, in which it subsequently cleaves POMC to the melanocortin peptides. This would suggest that alpha-MSH and beta-MSH may be subject to alternative sorting mechanisms, leading to heterogeneity in secretory granule content in POMC-producing cells. Overall, these studies implicate POMC processing as a key regulatory mechanism in the control of energy homeostasis.

Agouti-related protein (AGRP) plays a key role in energy homeostasis. The carboxyl-terminal domain of AGRP acts as an endogenous antagonist of the melanocortin-4 receptor (MC4-R). It has been suggested that the amino-terminal domain of AGRP binds to syndecan-3, thereby modulating the effects of carboxyl-terminal AGRP at the MC4-R. This model assumes that AGRP is secreted as a full-length peptide. In this study we found that AGRP is processed intracellularly after Arg(79)-Glu(80)-Pro(81)-Arg(82). The processing site suggests cleavage by proprotein convertases (PCs). RNA interference and overexpression experiments showed that PC1/3 is primarily responsible for cleavage in vitro, although both PC2 and PC5/6A can also process AGRP. Dual in situ hybridization demonstrated that PC1/3 is expressed in AGRP neurons in the rat hypothalamus. Moreover, hypothalamic extracts from PC1-null mice contained 3.3-fold more unprocessed full-length AGRP, compared with wild-type mice, based on combined HPLC and RIA analysis, demonstrating that PC1/3 plays a role in AGRP cleavage in vivo. We also found that AGRP(83-132) is more potent an antagonist than full-length AGRP, based on cAMP reporter assays, suggesting that posttranslational cleavage is required to potentiate the effect of AGRP at the MC4-R. Because AGRP is cleaved into distinct amino-terminal and carboxyl-terminal peptides, we tested whether amino-terminal peptides modulate food intake. However, intracerebroventricular injection of rat AGRP(25-47) and AGRP(50-80) had no effect on body weight, food intake, or core body temperature. Because AGRP is cleaved before secretion, syndecan-3 must influence food intake independently of the MC4-R.