Developmental Biology of Mammalian Melanocytes

0022-202X/ 80/7501-0116$02.00/0 THE JouRNAL OF INVESTIGATIVE DERMATOLOGY, 75:116-120, 1980 Copyright 1980 by The Williams & Wilkins Co. Vol. 75, No.1 Printed in U.S. A. Developmental Biology of Mammalian Melanocytes WALTER C. QuEVEDO, JR., PH.D., AND RoBERT D. FLEISCHMANN, PH.D. Division of Biology and Medicine, Brown University, Providence, Rhode Island U. S. A. The developmental biology of mammalian cutaneous melanocytes spans the entire life cycle from embryo to senescent adult. This review focuses on the melanocytes of the epidermis. Recent studies on neonatal mice indicate that melanoblasts of neural crest origin differentiate into epidermal melanocytes in response to signals provided by the local tissue environment, which may incorporate melanocyte-stimulating hormone and other hormones of pituitary origin in its action. In mice, there is a rapid disappearance of active epidermal melanocytes from the haired general body (trunk) skin during the 1st mo of life. Since active epidermal melanocytes persist in the relatively hairless skin of the appendages, some action of the local tissue environment is indicated here as well. An age-dependent decline in epidermal melanocytes occurs in humans over a period of many years. The fate of these melanocytes and the cause of their decline in humans and mice are not known. Several lines of evidence indicate that at least some melanocytes persist in an inactive form and may be stimulated to produce melanin on exposure to ultraviolet light. The local tissue environment influences not only the ability of epidermal melanocytes to synthesize melanin but also the types of pigment-bearing bodies (melanosomes) that they manufacture. In recessive yellow mice, for example, pheomelanosome-like bodies are synthesized by epidermal melanocytes whereas those of the dermis produce eumelanosomes. The changes observed during the clinical course of human piebaldism and vitiligo indicate that studies on age-dependent modifications in normal mouse and human epidermal melanocytes are essential for an understanding of these and other pigmentary disorders. Mintz has presented strong evidence that the entire melanocyte population of mouse skin is derived from 34 clonal initiator cells, each of which originates in the neural crest and migrates and proliferates to give a band of melanocytes that extends laterally (see Silvers [l] for references and an excellent summary and analysis of this work). The commitment of cells to form melanoblasts evidently occurs within the neural crest before migration. The genes responsible for melanoblast determination fu-st become active between days 5 and 7 of gestation; melanoblasts migrate to ail parts of skin between days 8.5 and 12 of embryonic age [2]. Melanoblasts enter skin via the dermis, migrate there, and secondarily invade the epidermis. It appears that melanoblasts penetrate the epidermis between the 11th and 12th days of development. Although diversity may exist among neural crest cells prior to migration, further developmental events are necessary for their differentiation into specialized cells. Thus, the intermediate environments through which neural crest cells migrate may influence their course of differentiation; in particular, the tissue environment in which Reprint requests to: Dr. Walter C. Quevedo, Jr., Division of Biology and Medicine, Box G, Brown University, Providence, Rhode Island 02912. This study was supported in part by PHS Grant #5-S07-RR07085. Abbreviations: camp: Cyclic adenosine 5' -monophosphate MSH: melanocyte-stimulating hormone UVL: ultraviolet light they become localized plays a major role in this process [3]. The implications of these findings for the origin of "white spotting" of skin and hair have been reviewed elsewhere [1]. Within a given tissue environment, melanocyte performance may change throughout life. Our current understanding of the developmental origin of the postnatal changes in melanocyte form and function provides the basis for this paper, which emphasizes data on the mouse and the implications of such data for man. DEVELOPMENT OF EPIDERMAL MELANOCYTES IN NEONATAL MICE Although in adult pigmented mice melanin-bearing melanocytes are present in the epidermis, dermis, and hair bulbs of relatively hairless skin of the feet, ears, tail, and scrotum, they are usually restticted to hair bulbs in hairy general body (trunk) skin. However, dopa-positive melanocytes are numerous in the epidermis of trunk skin at birth and reach a maximum population density (estimates range from 400 to l,ooo/mm 2 ) between days 2 and 4 postpartum [ 4-6]. They then gradually decline in number until only a few melanocytes remain on day 30. Use of a combined dopa-premelanin treatment to visualize epidermal melanocytes in C57BL (black) mice has shown that a constant number of melanocytes (approximately 1,400/mm 2 ) is mamtained from birth until day 4, after which the number declines [6]. This observation indicates that the method used to demonstrate epidermal melanocytes for counting is extremely important. The dopa-premelanin reaction developed by Mishima consists of the treatment of tissues incubated in dopa with ammoniated s~ver nitrate [7]. Dopa reveals tyrosinase activity, and ammo mated silver nitrate binds to melanin and elements of the melanosome ("premelanin") unrelated to active tyrosinase [7-9]. Thus, ammoniated silver nitrate is said to reveal tyrosinasenegative melanocytes (melanoblasts) containing unmelanized melanosomes as well as tyrosinase-positive melanocytes containing melanized or unmelanized melanosomes [7-9]. At the light microscopic level, then, dopa reveals melanocytes with active tyrosmase whereas dopa-premelanin reveals the larger population of melanosome-containing tyrosinase-positive and -negative melanocytes [6]. Treatment of tissues with dopa prior to ammoniated silver nitrate increases the probability that the latter will reveal a reactiv.e substrate in melanocytes. Mishima and Widlan [8] have demonstrated that greater numbers of epidermal melanocytes are identified in human skin by quantitative electron microscopy than by any available cytochemical method coupled with light microscopy. This fact suggests that owing to physiological variations in melanocytes, only a fraction of the population is evident at the light microscopic level regardless of the cytochemical test employed. The developmental mechanism triggering differentiation of epidermal melanoblasts into melanocytes in mice is not known. Injections of a-melanocyte-stimulating hormone (a-msh) and dibutyryl cyclic adenosine 5'-monophosphate (dibutyryl camp) into C57BL mice on days 1 and 2 post partum increase the numbers of dopa-positive melanocytes found 24 hr later [5,6]. There is no increase, however, in the numbers of melanocytes revealed by dopa-premelanin treatment. a-melanocyte-stinmlating hormone administered on days 19 and 20 post partum slightly increases the small number of dopa-positive melanocytes from approximately 3/.1 mm 2 to 7/.1 mm 2 [6]. The 116

July 1980 MELANOCYTE BIOLOGY 117 number of dopa-and-premelanin-positive cells is not increased. These facts have been interpreted to mean that a-msh and dibutyryl camp do not stimulate melanocyte proliferation but rather induce tyrosinase activity within melanocytes that otherwise would be dopa-negative [6]. On the basis of their studies, Hirobe and Takeuchi have suggested that a-msh may normally act in newborn mice to stimulate melanoblast differentiation into melanocytes by inducing tyrosinase activity, melanosome formation, translocation of melanosomes, and increased dendritogenesis [5,6). Although the time of onset of MSH release by the pituitary has not been established for the mouse, results of cytoimmunologic studies [10] have indicated that in rats synthesis of a-and {3- MSH apparently is under way by day 19 of gestation. At present, a-msh and,8-msh have not been clearly demonstrated in human pituitaries, and {3-MSH activity has been attributed to {3- and y-lipotropins containing the entire {3-MSH amino acid sequence [11,12]. Accordingly, the suggestion of a role for a-msh in the normal differentiation of mouse melanoblasts [5,6] seems to indicate that,8-msh, lipotropins, and ACTH (which incorporates the a-msh sequence) also influence the development of melanoblasts in mice [11]. It seems unlikely that MSH or an MSH-like compound is the sole agent responsible for melanoblast differentiation in the mouse because me.lanocytes develop within explants of embryonic skin in the absence of added hormones [3]. A major question yet to be answered is why just after appearing melanocytes disappear from the epidermis and dermis of the trunk during the first 4 wk of life [ 4,6). A specific age-dependent control mechanism in the local tissue envn onment is indicated by the fact that although extrafollicular melanotic melanocytes disappear from the trunk they persist in the epidermis of relatively hairless skin of the extremities. Melanocyte-stimulating hormone and dibutyryl camp are not the only agents found to increase the populations of active epidermal melanocytes in young mice. Ultraviolet light (UVL) also dramatically increased the number of dopa-positive epidermal melanocytes when administered to pigmented mice for 6 days beginning with the 16th day post partum; an average of less than 50/mm 2 were present on day 16, and 500/ mm 2 were present on day 22, a time when fewer than 50/mm 2 were present in nonirradiated trunk skin [4]. Since the combined dopa-premelanin method was not used in this study, the extent to which the increase in dopa-positive melanocytes reflected reactivation of melanocytes or their proliferation was not determined. Electron microscopic studies have helped clarify the cytologic events associated with the appearance and disappearance of epidermal melanocytes from trunk skin. Melanocytes containing developing melanosomes are present in the epidermis of trunk skin on day 15 of gestation and subsequently increase in number [13]. Numerous melanocytes containing developing and fully formed melanosomes are present until day 3 postpartum, after which they decline in number [13). Melanocytes present on day 6 and thereafter (Fig 1) appear to be fully mature in that they contain varying numbers of uniformly melanized melanosomes and few melanosomes in formative stages [14]. At this time some melanosomes are also present within keratinocytes. Melanocytes are found in progressively smaller numbers through the 18th day [14]. The fate of epidermal melanocytes is not clear. In one study, no definite evidence of melanocyte degeneration or uprooting was observed [14]. It was tentatively suggested that the epidermal melanocytes progressively ceased melanogenesis and became "amelanotic" when then melanosomes were completely transferred to receptor keratinocytes. It was proposed that at least some of the indeterminate cells present in mouse epidermis from bnth (Fig 2) were of the melanocyte lineage and that melanocytes joined this population upon losing then melanosomes [14]. On the other hand, probably at least some epidermal melanocytes were lost, for Weiss and Zelickson [13] found fragments of melanocytes within heterophagocytic vesicles of keratinocytes after day 4 post FIG l. Fully melanized melanosomes in epidermal melanocyte of an 8-day-old black (C57BL/ 6) mouse (X 14,000). FIG 2. Indeterminate epidermal ceu in an 18-day-old black (C57BL/ 6) mouse (reduced from x 14,000). partum. Some variation in the time of appearance and disappearance of melanocytes within the epidermis may have been related to the limited sampling possible at the electron microscopic level. ORIGIN OF MELANOGENIC MELANOCYTES IN NONPIGMENTED SKIN IRRADIATED WITH ULTRAVIOLET LIGHT Although exposure of depilated hairy and hairless pigmented adult mice to UVL results in a striking increase in dopa-positive

118 QUEVEDO AND FLEISCHMANN melanocytes within the trunk epidermis, the source of the melanocytes is not clear [15,16]. Recently, Miyazaki et al have identified, among the indeterminate cells of trunk epidermis, a population of cells that, on the basis of its location, response to UVL, and lack of response to 2,4-dinitrochlorobenzene, they consider to be amelanogenic melanocytes (17]. In their view, the increased numbers of active melanocytes after exposure of black mice to UVL involves activation of preexisting epidermal melanocytes as well as proliferation (17]. Rosdahl and Szabo [18] have provided evidence that essentially all of the increase in epidermal melanocytes in ear skin of black mice after exposme to UVL results from proliferation of melanotic melanocytes. They found a 4- to 6-fold increase in dopa-positive melanocytes in UVL-i.rradiated ear skin (e.g., a rise fro m 40 to 150/mm 2 ); all new melanocytes had taken up tritiated thymidine, an indication they resulted from the proliferation of the small native population of melanotic epidermal melanocytes. Rosdahl [19) has demonstrated mitotic activity within noni.rradiated pigment-producing melanocytes and has concluded that this is further evidence that these cells exhibit increased mitotic activity in response to UVL. Uesugi et al [20] have detected indeterminate cells containing 100-A (intermediate) filaments in the trunk epidermis of adult black mice. Serial biopsy specimens removed on days 0 to 5 of UVL exposw e indicated that they were the precursors of the melanotic melanocytes that appeared in UVL-irradiated skin. Their evidence supports the view that UVL-induced hyperpigmentation of normally nonpigmented trunk skin involves activation of precursor melanocytes as well as proliferation. Accordingly, in the ear pinna, the stem population of melanocytes induced to proliferate by UVL appears to be melanogenically active prior to irradiation, whereas in trunk skin it apparently is melanogenically inactive. In mice the number of melanocytes activated or proliferated ia response to UVL is influenced by major genes controlling coat coloration [15,16]. They influence the intensity of radiation-induced tanning by determining the size of melanocyte populations, melanocyte distribution, melanocyte structw e, the color and size of melanosomes, and th e total amount of pigment synthesized [1,15,16). EPIDERMAL MELANOCYTES IN HUMAN SKIN The problem of the fate of epidermal melanocytes in neonatal mice is not umelated to that of melanocyte disappearance from the human epidermis with advancing age. The epidermal melanocytes of nonexposed buttock, abdominal, and medial upper arm skin of Caucasians undergo a substantial reduction in their numbers throughout adult life (21,22). In the chronically sunexposed lateral aspect of the upper arm and in buttock skin exposed to sun lamps, the numbers of dopa-positive epidermal melanocytes are substantially greater than in corresponding nonirradiated skin for each age examined. The melanocyte populations of irradiated skin also show an age-related decline in size that closely parallels the pattern found in nonirradiated skin (21,22). On the basis of the great variation in dopa reactivity of nonirradiated melanocytes, it has been suggested that UVL elicits increased melanogenic activity in ma ny melanocytes that normally produce little or no melanin (8,21). Weakly melanogenic melanocytes may not darken demonstrably upon incubation in dopa and thereby escape detection in sheets of nonirradiated epidermis. The uniformly intense dopa reactivity of U:rad iated epidermal melanocytes gives the impression that the normally variable population has been stimulated to a comparable level of melanogenic activi ty (21). As discussed above, Mishima and Widlan [8] have concluded that the increase in numbers of melanocytes in human skin after UVL irradiation is the result of increased melanogenic activity within existing melanocytes as well as of other phenomena, possibly involving melanocyte proliferation. Gilchrest, Blog, and Szabo [22] have concluded that repeated sun exposure can irreversibly alter the size of melanocyte populations in the skin. Upper arm Vol. 75, No. 1 skin receiving no UVL for approximately 6 mo had substantially more melanocytes than unexposed skin. Gilchrest a nd co-workers have pointed out that the generalized hyperpigmentatipn found in habitually exposed skin of aged individuals occu1 s despite a reduction in melanocyte populations (22]. It ma:y be the result of greater tyrosinase activity within the surviving chronically irradiated melanocytes. At present it is not known whether the age-dependent decline in dopa-positive melanocytes reflects a true loss or simply conversion to an amelanotic form. EUMELANOSOMES, PHEOMELANOSOMES, AND EPIDERMAL MELANOCYTES In contrast to eumelanosomes, which are elliptical a nd have an organized fibrillar matrix (Fig 3), the pheomelanosomes of follicular melanocytes in dominant (lethal) and recessive yellow mice. consist primarily of micro vesicles surrounded by a limiting membrane (Fig 4 and Fig 5). Some disorganized filamentous elements may also be present [23,24]. Pheomelanin is deposited on the surfaces of the microvesicles and fil aments [23,24]. In agouti mice, a single melanocyte alternately synthesizes fibrillar eumelanosomes and microvesicular pheomelanosomes (23]. In the mouse, pheomelanosomes and eumelanosomes of follicular melanocytes (Fig 3, 4, a nd 5) at their eru liest stage of development are identical, i. e., spherical membrane-limited bodies with microvesicles and fllaments within their interior compartments. In pheomelanosomes the increased number of microvesicles gives rise to their chru acteristic matrix, whereas in eumelanosomes the filamentous component is amplified [24]. In pheomelanosomes, it is possible that tyrosinase is located inside the microvesicles and that it produces dopa quinone, which diffuses through the microvesicular membrane to combine with cysteine and give rise to cysteinyldopa, which is converted to pheomelanin. This untested hypothesis neatly explains how cysteine, a sulfhydl yl-group-containing amino acid, does not inhibit tyrosinase activity. The spotty deposition of pheomelanin on the fibrillru matrices of some human pheomelanosomes may indicate inhibitory effects of cysteine in direct contact with tyrosinase [25,26]. At the electron microscopic level it has been found that the epidermal melanocytes of black mice synthesize elliptical eumelanosmes chru acterized by fibrillar matrices with the typical periodicity of substructure in the individual fibrils.* Among yellow mice, recessive yellow has received the greatest attention thus far. In this genotype, the dermal melanocytes contain melanosomes that are elliptical in shape and that exhibit other features typical of eumelanosomes, i.e., melanosomes containing black-brown melanin. In nonirradiated recessive yellow mice, the melanosomes within dermal melanocytes apperu to be fully melanized, and no in1mature stages ru e evident. After exposure to UVL, some immature melanosomes are present within th e dermal melanocytes, an indication that exposure to UVL stimulates melanogenesis. Melanosomes of recessive yellow dermal melanocytes correspond closely in structure to eumelanosomes synthesized by epidermal melanocytes of C57BL (black) mice. On the basis of light and electron microscopy, it seems definite that dermal melanocytes of recessive yellow mice synthesize eumelanin. Light microscope findings suggest the same for viable yellow.and dominant (lethal) yellow mice. Epidermal melanocytes of recessive yellow mice appear to synthesize melanosomes quite distinct from eumelanosomes of dermal melanocytes. The granules are round, membrane-limited, and contain a particle-studded interior (Fig 6). The electron density of the interior of these "melanosomes" vru ies among granules and between skin specimens, perhaps a reflection of differing degrees of mela nization. In overall appearance they resemble most closely pheomelanosomes within hair follicles of recessive yel- *Quevedo WC Jr, Fleischmann RD, Dyckman J, unpublished observations.

July 1980 MELANOCYTE BIOLOGY 119 dominant (lethal) yellow (AY/-) mice suggests that the m elanosomes contain vesicular and fibrillar structuxes. Theil overall appearance is different from that of melanosomes in epidermal melanocytes of recessive yellow mice. Under a light microscope, the small pigment granules of recessive yellow and dominant (lethal) yellow mice appear m ore brown th an yellow. This finding, coupled with observations on dermal melanocytes, has led to the tentative conclusion that the extrafollicular melanocytes of yellow mice synthesize eumelanin and that ph eomelanin is synthesized only by melanocytes wit hin haii follicles [1]. On the basis of similarities a mong melanosomes in human intern eckle melanocytes, melanosomes in recessive yellow epidermal m ela nocytes, and pheomelanosomes of follicular melanocytes in recessive yellow mice and in some red-haired people, the question of pheomelanin synth esis within extrafollicular melanocytes must remain open. FIG 3. Eumelanosomes in folli cular melanocyte of ad ul t black (C57BL/6) mouse (loth day of hair growth cycle). One melanosome is completely melanized; another is at an ea rly stage of melanin deposition on the "spindle-shaped" matrix (reduced from X 90,000). FIG 4. Pheomelanosome in follicular melanocyte of an adult recessive yellow (C57BL/ Ha-e ) mouse (loth day of hair growth cycle ). The rnelanosome contains a matrix of vesicles and filaments with varying amounts of melanin deposited on it (reduced from X 90,000). FIG 5. Pheomelanosome in follicular melanocyte of an adult recessive yellow (C57BL/ Ha-e ) mouse (loth day of hai.r growth cycle). The rnela.nosome contains a matrix of vesicles and filaments with varying amounts of melanin deposited on it (reduced from X 90,000). FIG 6. Melanocyte in the tail epidermis of a recessive yellow (C57BL/ Ha-e ) mouse treated with ultraviolet light (reduced from X 14,570). Inset, this melanosome, like those shown at a lower magnification, is closer in general structure to the pheomelanosomes of Fig 4 and Fig 5 than the eumelanosomes of Fig 3 (red uced from X 86,800). low mice (Fig 4 and Fig 5) and certain granular organelles found in melanocytes of the interfreckle epidermis of red-ha ired, freckled people [27). Pa.ralleling observations on the follicular melanocytes of recessive yellow mice [24] is the observation that occasional epidermal melanocytes contain, in addition to pheom elanosomes, a few melanosomes that resemble eumelanosom es. Preliminary examination of epidermal melanocytes of EXPERIMENTAL APPROACHES TO THE REGULATION OF MELANOBLAST DIFFERENTIATION Studies on the d e~elopm e n tal interactions involved in the differentiation of m elanocytes within skin of yellow mice illustrate th e complexity of this process. By means of skin transplantations that permitted mela noblasts of a given genotype to invade hai.r bulbs of another genotype in neonatal mice, Silvers and Russell [29] demonstrated that in dominant (lethal) yellow (AY I-) mice the haii follicle envii onment specifies t he production of pheomelanin. In a broader sense, they found that the genotype of the follicle, not that of the melanocyte, determines the production of pheomelanin regulated by agouti locus alleles (A- v, A vy, A, a, etc.). More specifically, Mayer and Fish bane [29] and Poole [30] separated epidermis and dermis of 13- to 17-day-old embryonic skin and found, upon recombining epidermis and dermis of differing "genotypes, t hat the alleles at the agouti locus act via the dermis. In general, the a-locus genotype of the epidermis plays no obvious role in determining the type of melanin synthesized. An exception is Av dominant (lethal) yellow; in this case both the epidermis and dermis specify the synthesis of yellow melanin by mela nocytes of hair follicles [30]. Thus, action of the agouti locus may originate in the dermis, a nd it is transduced via the epidermis. Only in the case of Av is the information of a long dmation. By recombining neural crest tissue and skin n om mouse embryos of el e and E + I E + genotypes, Lamoreux and M ayer [31] demonstrated that the e locus acts via the nuclear genome of the m elanocyte and not through the tissue environment. Recessive yellow (e/e) melanocytes req uire a hair follicle in which to produce pheomelanin, but apparently a haii follicle of any genotype suffices [31]. Wild type (E+ I E +) melanocytes invading el e haii follicles produce eumelanin, whereas el e melanocytes synthesize pheomelanin within E + I E + hair fo llicles. A similar conclusion was reached independently by Poole and Silvers, who used different techniques [32]. In view of the general similarity in the ultrastructure of mela nosomes in the follicular melanocytes of recessive yellow a nd lethal yellow mice, it might seem at first glance that the e and a loci influence different steps in a common pathway leading to the formation of mature pheomelanosomes. However, in view of the complexity of expression of a locus and e locus alleles, a mechanism involving t he production of inhibitors and counterinhibitors within t he hair follicle appears to explain the available data more rationally [33]. One po sible model is that thee+ allele specifies an inhibitor-of pheomelanin synthesis within melanocytes, while the A.v allele specifies a substance that neutralizes the influence of the e locus and results in pheomelanogenesis. In the presence of t he e allele, no inhibitor is produced within melanocytes and pheomelanogenesis proceeds without a requirement for action of alleles at the agouti locus, an action in melanogenesis that appears to take place only within the envii onment of the hair follicle [28).

120 QUEVEDO AND FLEISCHMANN The observation that melanosomes in epidermal melanocytes of recessive yellow mice are different from eumelanosomes within dermal melanocytes of the same animal indicates that each skin region places its own stamp on melanocyte function, and, as indicated above, that such influences may change with age. IMPLICATIONS FOR HUMAN PIEBALDISM The changing nature of the epidermis/dermis in influencing the function of epidermal melanocytes is evident in some people with piebaldism. Melanocytes are absent from the epidermis and hair follicles of the characteristic white forelock, but are present in the hypermelanotic islands within the hypopigmented patches of skin of the trunk and extremities [34]. Melanocytes are rarely found in the hypopigmented epidermis surrounding these hypermelanotic "islands." In both regions they produce abnormal melanosomes [34]. The evidence suggests that melanoblasts spread from the neural crest to all regions of the skin and either fail to survive or differentiate in many regions. Where they do differentiate in hypopigmented areas, they function abnormally [34]. In a recent study on presumed human piebaldism, it was found that the white forelock diminished in size with advancing age but the hypopigmented areas of the trunk and extremities became more extensive [35]. This finding suggests that at least in some piebald people there are postnatal gains and losses in the ability of specific regions of skin to support the function of differentiated melanocytes. A generally similar statement may be made for the clinical course of vitiligo. CONCLUSION This brief review has provided further documentation that there is a close interplay between studies on the biology of melanocytes in mice and those on the origin and progression of age-dependent melanocyte loss, piebaldism, and vitiligo in man. Knowledge of these processes in mice will invariably catalyze an understanding of parallel developmental events in man. REFERENCES 1. Silvers WK: The Coat Colors of Mice: A Model for Mammalian Gene Action and Interaction. New York, Springer-Verlag, 1979, pp 379 2. Mayer TC: The migratory pathway of neural crest cells into the skin of mouse embryos. Dev Bioi 34:39-46, 1973 3. Mayer TC, Oddis L: Pigment cell differentiation in embryonic mouse skin and isolated epidermis: An in vitro study. J Exp Zoo! 202:415-424, 1977 4. Quevedo WC Jr, Youle MC, Rovee DT, Bienieki TC: The developmental fate of melanocytes in mw ine skin, Structure and control of the Melanocyte. Edited by G Della Porta, 0 Muhlbock. Berlin, Springer-Verlag, 1966, pp 228-241 5. Hirobe T: Stimulation of dendritogenesis in the epidermal melanocytes of newborn mice by melanocyte-stimulating hormone. J Cell Sci 33:371-383, 1978 6. Hirobe T, Takeuchi T: Induction of melanogenesis in the epidermal melanoblasts of newborn mouse skin by MSH. J Embryo! Exp Morpho! 37:79-90, 1977 7. Mishima Y: New technic for comprehensive demonstration of melanin, premelanin, and tyrosinase sites. J Invest Dermatol34:355-360, 1960 8. Mishima Y, Widlan S: Enzymically active and inactive melanocyte populations and ultraviolet irradiation: Combined dopa-premelanin reaction and electron microscopy. J Invest Dermatol 49: 273-281, 1967 9. Mishima Y: Electron microscopic studies on the process of melan- Vol. 75, No. 1 ization of pigment cells in vivo and in vitro. Excerpta Medica International Congress Series 55:1217-1222, 1962 10. Dupouy JP, Dubois MP: Ontogenesis of the a-msh,,8-msh and ACTH cells in the foetal hypophysis of the rat. Correlation with the growth of the adrenals and adrenocortical activity. Cell Tissue Res 161:373-384, 1975 11. Gilkes JJH: The pituitary mela notrophic hormones. Br J Dermatol 91:353-355, 1974 12. Brown JD, Doe RP: Pituitary pigmentary hormones; relationship of melanocyte-stimulating hormone to lipotrophic hormone. JAMA 240:1273-1278, 1978 13. Weiss LW, Zelickson AS: Dendritic cells in developing mammalian epidermis. J Invest Derma to) 58:171, 1972 14. Quevedo WC Jr, Smith J: E lectron microscope observations on the postnatal "loss" of interfollicular epidermal melanocytes in mice. J Cell Bioi 39:108a, 1968 15. Quevedo WC Jr, Smith JA: Studies on radiation-induced tanning of skin. Ann NY Acad Sci 100:364-388, 1963 16. Quevedo WC Jr, McTague CF: Genetic influences on the response of mouse melanocytes to ultraviolet light: The melanocyte system of hair-covered skin. J Exp Zoot 152:159-168, 1963 17. Miyazaki H, Kawada A, Taka ki Y, Sato T, Masutani M: Effects of ul traviolet light on epidermal dendritic cells of hai..rless mice Sunlight and Man. Edited by TB Fitzpatrick, MA Pathak, LC Harber, M Seiji, A Kukita. Tokyo, University of Tokyo Press, 1974, pp 217-229 18. Rosdahl IK, Szabo G: Mitotic activity of epidermal melanocytes in UV-irradiated mouse skin. J Invest Dermatol 70:143-148, 1978 19. Rosdahl IK: Melanocyte mitosis in UVB-irradiated mouse skin. Acta Dermato Venereol (Stockh) 58:217-221, 1978 20. Uesugi T, Mitsuko K, Horikoshi T, Sadao S, Jimbow K: Mode of activation and differentiation of dormant melanocytes after UV exposure of mouse skin, Pigment Cell, vol 4, Biologic Basis of Pigmentation. Edited by SN Klaus. Basel, S Karger, 1979, pp 337-344 21. Quevedo WC Jr, Szabo G, Virks J: Influence of age and UV on the population of dopa-positive melanocytes in human skin. J Invest Dermatol 52:287-290, 1969 22. Gilchrest BA, Blog FB, Szabo G: Effects of aging and chronic sun exposure on melanocytes in human skin. J Invest Dermatol 73: 141-143, 1979 23. Sakurai T, Ochiai H, Takeuchi T: Ultrastructural change of melanosomes associated with agouti pattern J'OJ"mation in mouse hair. Dev Bioi 47:466-471, 1975 24. Fleischmann RD: Similarities and contrasts in the synthesis of pheomelanosomes and eumelanosomes in mouse follicular melanocytes. Yale J Bioi Med 50:556, 1977 (abstract #38) 25. Stanka P: Ultrastructural study of pigment cells of human red hair. Cell Tissue Res 150:167-178, 1974 26. Mottaz JH, Zelickson AS: Ultrastructu1 e of hai..r pigment, Advances in Biology of Skin, vol 9, Hai..r Growth. Edited by W Montagna, RL Dobson. Oxford, Pergamon Press, 1969, pp 471-489 27. Breathnach AS: E lectron microscopy of melanocytes and melanosomes in freckled human epidermis. J Invest Dermatol 42:339-394, 1964 28. Silvers WK, Russell ES: An experimental approach to action of genes at the agouti locus in the mouse. J Exp Zool 130:199-220, 1955 29. Mayer TC, Fishbane JL: Mesoderm-ectoderm interaction in the production of the agouti pigmentation pattern in mice. Genetics 71:297-303, 1972 30. Poole TW: Dermal-epidermal interactions and the action of alleles at the agouti locus in the mouse. Dev Bioi 42:203-210, 1975 31. Lamoreux ML, Mayer TC: Site of gene action in the development of hair pigment in recessive yellow (e/e) mice. Dev Biol46:160-166, 1975 32. Poole TW, Silvers WK: An experimental analysis of the recessive yellow coat color mutant i..n the mouse. Dev Bioi 48:377-381, 1976 33. Wolff GL, Galbraith DB, Domon OE, Row JM: Phaeomelanin synthesis and obesity in mice. J Hered 69:295-298, 1978 34. Jimbow K, Fitzpatrick TB, Szabo G, Hori Y: Congenital circumscribed hypomelanosis: A characterization based on electron microscopic study of tuberous sclerosis, nevus depigmentosus, and piebaldism. J Invest Dermatol 64:50-62, 1975 35. Davis BK, Verdol LD: Expansion and contraction of hypomelanotic areas i..n human piebaldism. Hum Genet 34:163-170, 1976