Serum-free culture of normal human melanocytes: growth kinetics and growth factor requirements

Normal human epidermal melanocytes were selectively propagated from mixed (keratinocyte-melanocyte) cultures and primary epidermal cell suspensions in serum-free medium, MCDB 153 containing insulin, bovine pituitary extract (BPE), phorbol-12-myristate-13-acetate (PMA), ethanolamine, phosphoethanolamine, and hydrocortisone. Neonatal foreskin melanocytes (NFMs) replicated more readily than adult melanocytes in culture. Early passage NFMs grown in serum-free medium exhibited a population generation time of 24-48 hours. NFMs assumed a less dendritic appearance and were less pigmented than adult melanocytes. PMA or other protein kinase C-activating phorbol esters significantly enhanced mitogenesis of NFMs; however, cAMP-elevating agents were not required for efficient replication of NFMs. Basic fibroblast growth factor (bFGF) was a potent mitogen for NFMs and replaced the requirement for BPE in the culture medium. NFMs expressed a single class of specific, high-affinity receptors for bFGF, exhibiting a Kd = 3 x 10(-11) M and approximately 76,500 receptors/cell. Neither EGF nor TGF-alpha were mitogenic for NFMs, and TGF-beta reversibly inhibited NFM growth. Rapidly growing, early passage NFMs were shown to have cell cycle times of 19.5, 7.5, and 9 hours for G1, S, and G2/M phases of the cell cycle, respectively. Culture of NFMs to confluence or depletion of growth factors from the culture medium caused reversible, G1 phase-specific, cell cycle growth arrest. Senescence of NFMs was associated with irreversible growth arrest in the G1 phase after 40-45 population doublings in culture. Our data demonstrate that basal medium MCDB 153 can be supplemented with defined factors to cultivate selectively two major constituent cell types of the epidermis, the melanocyte and the keratinocyte.     

Role of leukemia inhibitory factor in the regulation of the proliferation and differentiation of neonatal mouse epidermal melanocytes in culture

Mouse epidermal melanoblasts/melanocytes preferentially proliferated from disaggregated epidermal cell suspensions derived from newborn mouse skin in a serum-free melanoblast/melanocyte-proliferation medium supplemented with dibutyryl adenosine 3':5'-cyclic monophosphate (DBcAMP) and/or basic fibroblast growth factor (bFGF). Leukemia inhibitory factor (LIF) supplemented to the medium from initiation of primary culture increased the proliferation of melanoblasts or melanocytes as well as the differentiation of melanocytes. Pure cultured primary melanoblasts or melanocytes were further cultured with the medium supplemented with LIF from 14 days (keratinocyte depletion). LIF stimulated the proliferation of melanoblasts or melanocytes as well as the differentiation of melanocytes in the absence of keratinocytes. Moreover, anti-LIF antibody supplemented to the medium from initiation of primary culture inhibited the proliferation of melanoblasts or melanocytes as well as the differentiation of melanocytes. These results suggest that LIF is one of the keratinocyte-derived factors involved in regulating the proliferation and differentiation of neonatal mouse epidermal melanocytes in culture in cooperation with cAMP elevator and bFGF.     

Hepatocyte growth factor controls the proliferation of cultured epidermal melanoblasts and melanocytes from newborn mice

Mouse epidermal melanoblasts and melanocytes preferentially proliferated from disaggregated epidermal cell suspensions derived from newborn mouse skin in a serum-free melanocyte-proliferation medium (MDMD) and melanoblast-proliferation medium (MDMDF) supplemented with dibutyryl adenosine 3':5'-cyclic monophosphate (DBcAMP) and/or basic fibroblast growth factor (bFGF). Pure cultured primary melanoblasts and melanocytes were further cultured with MDMD/MDMDF supplemented with hepatocyte growth factor (HGF) from 14 days (keratinocyte depletion). The HGF increased the number of melanoblasts and melanocytes, but not the percentage of differentiated melanocytes in the melanoblast-melanocyte population in the absence of keratinocytes. Flow cytometry analysis showed that melanoblasts and melanocytes in the S and/or G2/M phases of the cell cycle were increased by the treatment with HGF. Moreover, an anti-HGF antibody supplemented to MDMD/MDMDF from the initiation of the primary culture (in the presence of keratinocytes) inhibited the proliferation of melanoblasts and melanocytes, but not the differentiation of melanocytes. These results suggest that HGF is a keratinocyte-derived factor involved in regulating the proliferation of epidermal melanoblasts and melanocytes from newborn mice in cooperation with cAMP elevators and/or bFGF.     

Basic fibroblast growth factor requires a long-lasting activation of protein kinase C to induce cell proliferation in transformed fetal bovine aortic endothelial cells.

Basic fibroblast growth factor (bFGF) induces a protein kinase C (PKC)-dependent mitogenic response in transformed fetal bovine aortic endothelial GM 7373 cells. A long-lasting interaction of bFGF with the cell is required to induce cell proliferation. bFGF-treated cells are in fact committed to proliferate only after they have entered the phase S of the cell cycle, 12-14 h after the beginning of bFGF treatment. Before that time, the mitogenic response to bFGF is abolished by 1) removal of extracellular bFGF by suramin, 2) addition of neutralizing anti-bFGF antibodies to the culture medium, 3) inhibition of PKC activity by the protein kinase inhibitor H-7, and 4) down-regulation of PKC by cotreatment with phorbol ester. Thus the requirement for a prolonged interaction of bFGF with the cell reflects the requirement for a prolonged activation of PKC. Similar conclusions can be drawn for the PKC activators 12-O-tetradecanoyl phorbol 13-acetate and 1,2-dioctanoyl-sn-glycerol. The two molecules require 16 and 6 h, respectively, of activation of PKC to induce 50% of maximal cell proliferation. The requirement for a long-lasting activation of PKC appears to be a mechanism for the control of cell proliferation capable of discriminating among transient nonmitogenic stimuli and long-lasting mitogenic stimuli.     

Multiple Pathways of N-Kinase Activation in PC12 Cells

Past work established a cell-free assay for a nerve growth factor (NGF)-activated protein kinase activity (designated N-kinase) that utilizes tyrosine hydroxylase and histone H1 as substrates and that is distinct from a variety of well-characterized kinases. This study explores the specificity and mechanistic pathway(s) by which N-kinase activity is regulated in PC12 rat pheochromocytoma cells. N-kinase is rapidly activated in these cells by treatment with NGF, epidermal growth factor (EGF), basic fibroblast growth factor (bFGF), phorbol ester, or dibutyryl cyclic AMP. Our data indicate that the stimulated activity is the same for each agent by several criteria: It exhibits the same characteristic biphasic elution pattern by Mono S fast protein liquid chromatography (FPLC), except for the case of dibutyryl cyclic AMP in which one of the activity peaks is somewhat shifted; it shows the same elution pattern by FPLC on a Superose 12 column; it possesses identical substrate specificity; and, except in the case of dibutyryl cyclic AMP, it does not show additivity when each agent is added simultaneously with NGF. The multiple forms of N-kinase are interconvertible in that rechromatography on a Mono S column yields a single peak of activity. Also, when NGF and dibutyryl cyclic AMP are simultaneously presented to cells, the chromatographic profile resembles that with NGF alone. Activation occurs through several independent initial pathways. Down-regulation of protein kinase C by phorbol ester pretreatment prevents N-kinase activation by phorbol ester, but not by the other agents. A PC12 cell-derived line deficient in cyclic AMP-dependent protein kinase II activity exhibits N-kinase activation by all treatments except dibutyryl cyclic AMP. The properties of N-kinase suggests that it is similar or identical to the ribosomal S6 protein kinase described by Blenis and Erikson. Additional experiments revealed that N-kinase activity can be stimulated in several cell lines in addition to PC12 cells. These findings indicate that the N-kinase can be activated via multiple second-messenger pathways and that it could therefore potentially play a significant role in mediating shared intracellular responses to various extracellular signals.     

Effects of commonly used mitogens on the cytotoxicity of 4-tertiary butylphenol to human melanocytes

Summary In the search for environmental compounds responsible for contact or occupational vitiligo, it was found that the most potent was 4-tertiary butylphenol (4-TBP). Exposure to 4-TBP is widespread both in industry and in consumer items including synthetic leather, plastic, glues, and germicidal phenolic detergents. How 4-TBP causes depigmentation and the death of melanocytes is currently unclear. Growth mitogens for human melanocytes include α-melanocyte stimulating hormone (α-MSH), basic fibroblast growth factor (bFGF) and 12-o-tetradecanoylphorbol-13-acetate (TPA). The former two mitogens are physiological growth factors for melanocytes. We have studied the effects of these mitogens on the cytotoxicity of 4-TBP in human melanocytes. Our results demonstrated that deprivation of α-MSH or bFGF from melanocyte cultures resulted in reduced cytotoxicity to 4-TBP. Similar results were obtained upon treatment of melanocytes with an inhibitor of cAMP-dependent protein kinase A (PKA), that is known to be activated by α-MSH, or with an inhibitor of the tyrosine kinase bFGF receptor. In contrast, removal of fetal bovine serum or TPA from the culture medium did not influence the susceptibility of melanocytes to 4-TBP. These results suggest that activation of the cAMP and tyrosine kinase signaling pathways, both of which are involved in the mitogenic response of melanocytes, increase the susceptibility of these cells to the cytotoxic effects of 4-TBP. This work was presented in part at the seventh meeting of Pan-American Society for Pigment Cell Research, June 15–18, 1997, Providence, Rhode Island, USA.     

Steel factor controls the proliferation and differentiation of neonatal mouse epidermal melanocytes in culture

Mouse epidermal melanoblasts and melanocytes preferentially proliferated from disaggregated epidermal cell suspensions derived from newborn mouse skin in a serum-free melanocyte-proliferation medium (MDMD) and melanoblast-proliferation medium (MDMDF) supplemented with dibutyryl adenosine 3':5'-cyclic monophosphate (DBcAMP) and/or basic fibroblast growth factor (bFGF). Pure cultured primary melanoblasts and melanocytes were then further cultured with MDMD/MDMDF supplemented with steel factor (SLF) (keratinocyte depletion). SLF increased the number of melanoblasts and melanocytes as well as the proportion of differentiated melanocytes in the absence of keratinocytes. Flow cytometric analysis showed that melanoblasts and melanocytes in the S and G2/M phases of the cell cycle were increased by treatment with SLF. Moreover, an anti-SLF antibody added to MDMD/MDMDF from the initiation of the primary culture (in the presence of keratinocytes) inhibited the proliferation of melanoblasts and melanocytes as well as the differentiation of melanocytes. These results suggest that SLF is one of the keratinocyte-derived factors involved in regulating the proliferation and differentiation of neonatal mouse epidermal melanocytes in culture in cooperation with cAMP elevator and bFGF.     

A model for studying the transmission of information produced by certain growth factors: activation mechanisms of S6 kinase in cultured astrocytes

Treatment of cultured astrocytes from 2-day-old rat cerebral hemispheres with insulin, somatomedin C (IGF1), thrombin and acidic or basic fibroblast growth factors promoted a rapid activation of a cytosolic protein kinase (S6 kinase) which phosphorylates ribosomal protein S6. The phorbol ester (TPA) also triggered a rapid increase in S6 kinase activity. Two agonists of adenylate cyclase activity (forskolin and isoproterenol) and the cyclic AMP analog (dibutyryl cAMP) also stimulated the same S6 kinase. These observations support the idea that several pathways might promote the activation of the same entity that is regarded as one of the primary targets of signals elicited by growth factors.     

Serotonin-induced deoxyribonucleic acid synthesis in vascular smooth muscle cells involves a novel, pertussis toxin-sensitive pathway

Serotonin-induced DNA synthesis in bovine aortic smooth muscle cells was totally abolished by pretreatment of cultures with 5 ng/ml pertussis toxin. The half maximally effective concentration of toxin was approximately 10 pg/ml. Pertussis toxin did not affect platelet-derived growth factor (PDGF)-stimulated DNA synthesis and actually enhanced the mitogenic effect of the phorbol ester, phorbol 12-myristate 13-acetate. Pertussis toxin did not inhibit serotonin-stimulated inositol phosphate accumulation or increases in intracellular calcium or cAMP concentrations under conditions sufficient to completely inhibit serotonin-induced (3H)thymidine incorporation. These results demonstrate that a novel, pertussis-sensitive pathway is required for serotonin-, but not platelet-derived growth factor-induced DNA synthesis in vascular smooth muscle cells. The pertussis-sensitive step does not involve cAMP, phosphoinositide hydrolysis, mobilization of intracellular calcium, or phorbol ester-sensitive protein kinase C activity.     

Synergistic stimulation of DNA synthesis by cyclic AMP derivatives and growth factors in mouse 3T3 cells

Addition of the cAMP derivatives butcAMP or 8BrcAMP to quiescent cultures of Swiss 3T3 causes synergistic stimulation of DNAk synthesis with insulin, phorbol esters, vasopressin, epidermal growth factor, or fetal bovine serum (2-5%). In the presence of insulin, 8BrcAMP, and butcAMP stimulate [3H]-thymidine incorporation into acid-precipitable material in a dose-dependent manner. The effect of these agents is specific since 8Br5'AMP, 5'AMP, butyrate, or 8BrcGMP fail to stimulate DNA synthesis under identical experimental conditions. Furthermore, the mitogenic effects of the cAMP derivatives were markedly potentiated by 1-methyl-3-isobutyl xanthine and 4-(3-butoxy-4-methoxy benzyl)-2-imidazolidine, both of which are potent inhibitors of cyclic nucleotide phosphodiesterase activity. The growth-promoting effects of the cAMP derivatives were demonstrated by [3H]-thymidine incorporation (either by scintillation counting or by autoradiography), by flow cytofluorometric analysis, and by increase in cell number. When quiescent Swiss 3T3 cells were exposed to butcAMP and insulin, DNA synthesis began after a lag of 17h. The result of sequential additions of cAMP derivatives and insulin to quiescent 3T3 cells suggest that these agents must act simultaneously in G0/G1 to stimulate entry into DNA synthesis in these cells. The findings support the proposition that an increase in cellular levels of cAMP (but not cGMP) act sas a mitogenic stimulus for confluent and quiescent Swiss 3T3 cells.     

Na+-dependent elevation of the acidic cell surface pH (microclimate pH) of rat jejunal villus cells induced by cyclic nucleotides and phorbol ester: possible mediators of the regulation of the Na+/H+ antiporter

The effects of cyclic nucleotides and phorbol ester on the acidic cell surface pH of rat jejunal villi were studied by using single-barrelled pH-sensitive microelectrodes. Addition of dibutyryl cAMP (1 mM) to the mucosal bathing solution caused an elevation of the cell surface pH from 6.19 +/- 0.04 (n = 12 measurements from three animals) to 6.53 +/- 0.03 (12) in the presence of Na+ in the medium. However, dibutyryl cAMP had no significant effect in the absence of Na+ and presence of 1 mM amiloride. Dibutyryl cGMP (1 mM) also had an Na+-dependent inhibitory effect on the cell surface pH. A phorbol ester, phorbol 12-myristate 13-acetate, caused an elevation of the cell surface pH only in the presence of Na+ from 6.14 +/- 0.07 (12) to 6.46 +/- 0.08 (12). Phorbol and phorbol 13-acetate, which do not stimulate protein kinase C, were without significant effects. These results suggest that increased levels of the intracellular cyclic nucleotides and activation of protein kinase C raise the acidic cell surface pH by inhibiting the activity of the brush-border Na+/H+ antiporter in the rat jejunal villus cells.     

Granulocyte-macrophage colony-stimulating factor is a keratinocyte-derived factor involved in regulating the proliferation and differentiation of neonatal mouse epidermal melanocytes in culture

Mouse epidermal melanoblasts and melanocytes preferentially proliferated from disaggregated epidermal cell suspensions derived from newborn mouse skin in a serum-free melanocyte-proliferation medium (MDMD) and a melanoblast-proliferation medium (MDMDF) supplemented with dibutyryl adenosine 3':5'-cyclic monophosphate (DBcAMP) and/or basic fibroblast growth factor (bFGF). Pure cultured primary melanoblasts and melanocytes were further cultured with MDMD/MDMDF supplemented with granulocyte-macrophage colony-stimulating factor (GMCSF) from 14 days (keratinocyte depletion). GMCSF stimulated the number of melanoblasts/melanocytes as well as the percentage of differentiated melanocytes in keratinocyte-depleted cultures. Flow cytometry analysis showed that melanoblasts and melanocytes in the S and G(2)/M phases of the cell cycle were increased by the treatment with GMCSF. Moreover, anti-GMCSF antibody added to MDMD/MDMDF from the initiation of the primary culture (in the presence of keratinocytes) inhibited the proliferation of melanoblasts/melanocytes as well as the differentiation of melanocytes. Enzyme-linked immunosorbent assay of culture media revealed that GMCSF was secreted from keratinocytes, but not from melanocytes. These results suggest that GMCSF is one of the keratinocyte-derived factors involved in regulating the proliferation and differentiation of neonatal mouse epidermal melanoblasts/melanocytes in culture in cooperation with cAMP elevator and bFGF.     

Terminal differentiation and senescence in the human melanocyte: repression of tyrosine-phosphorylation of the extracellular signal-regulated kinase 2 selectively defines the two phenotypes.

Melanocytes are pigmented cells distributed in humans in several organs like the epidermis, the leptomeninges, the eye, and the inner ear. Epidermal melanocytes, whether derived from adult or neonatal skin, proliferate well in a medium supplemented with phorbol esters and other mitogens before they undergo senescence. Potent cAMP inducers like cholera toxin are also growth promoters for neonatal melanocytes but only transient growth stimulators for cells derived from adults. We used this cellular system to delineate biochemical pathways involved in proliferation and in terminal differentiation. Here we show that after a period of 4-8 wk of sustained proliferation in the presence of cholera toxin, the adult melanocytes became round, flat, and enlarged. These changes were associated with terminal growth and preceded by a five- to sixfold increase in cAMP levels and an 8- to 10-fold increase in melanin content. The simultaneous addition of phorbol esters and cholera toxin did not prevent cells from reaching terminal differentiation. Identified targets for phorbol esters are protein kinase C (PKC) and the mitogen-activated kinases (MAPKs), also called extracellular signal-regulated kinases (ERKs). PKC was found to be similarly regulated in proliferating and in terminally differentiated melanocytes. Proliferating melanocytes in early or late passage showed identical activation of the kinase ERK2. This kinase was rapidly phosphorylated upon phorbol 12-myristate 13-acetate (PMA) addition and specifically accumulated in the nucleus of the cells, whereas in unstimulated cells it had a perinuclear distribution. In contrast, senescent and terminally differentiated cells were unable to phosphorylate tyrosine residues of the ERK2 gene product in spite of presenting normal amounts of ERK2 protein. In addition, ERK2 did not show the nuclear accumulation observed in proliferating melanocytes after PMA activation and remained localized in the perinuclear area. These results demonstrate that senescent and terminally differentiated melanocytes share a common block in a critical pathway thought to integrate multiple intracellular signals transmitted by various second messengers and specifically prevent the continuation of the signal transduction cascade initiated by PMA activation of PKC.     

Modulation of Normal Human Melanocyte Dendricity by Growth-Promoting Agents

Dendrite formation and extension, which comprise a characteristic morphology of human normal melanocytes in the skin, represent one of the functional activities of melanocytes, the ability to transfer melanosomes into neighboring keratinocytes. However, the morphology of the melanocyte in vitro is usually quite different from that observed in vivo. it is probably due to the hyperproliferative condition of the melanocytes in culture. No studies have ever compared the effects of a single factor on both dendricity and proliferation at the same time. Therefore, we have compared the effects of six growth-promoting agents commonly used for melanocyte cultures on dendrite formation and proliferation. The addition of agents that increase the intracellular levels of cyclic adenosine monophosphate (cAMP)—dibutyryl cyclic adenosine monophosphate (db cAMP; 1 mM) or isobutylmethyl xanthine (IBMX; 0.1 mM)—had a strong effect on dendrite formation and a negative effect on proliferation. This was especially true with db cAMP. In the presence of 2% or 5% of heat-inactivated fetal bovine serum (FBS), dendrite formation was significantly increased as was proliferation. The number of dendrites was decreased in the culture with 12-o-tetradecanoylphorbol-13-acetate (TPA), but cell growth was slightly increased. With human recombinant basic fibroblast growth factor (bFGF) (0.5, 1.0 ng/ml) in the presence of bovine pituitary extract (BPE) (60 μg/ml), cell growth was increased. With 2 ng/ml of bFGF, however, a strong inhibitory effect on proliferation was observed. However, dendrite formation was constant at all concentrations of bFGF tested (0.5, 1.0 or 2.0 ng/ml) with BPE (30 or 60 μg/ml). In this study, we have demonstrated that dendrite formation was suppressed by the reagents that stimulate melanocyte proliferation, and vice versa, with the only exception being heat-inactivated FBS. Both dendrite formation and proliferation were induced by the heat-inactivated FBS. This approach is crucial to the development of an adequate culture system for proliferation and/or dendrite formation of normal human melanocytes. It is necessary to keep these aspects in mind as we further investigate the biology of melanocytes, especially the cell-to-cell interactions between melanocytes and keratinocytes, involved in melanogenesis and melanin pigmentation in vivo. This study also provides practical and important information for a future reconstitutive skin system composed of melanocytes, keratinocytes, and fibroblasts in a single culture medium.     

Basic FGF and TGF-beta 1 influence commitment to melanogenesis in neural crest-derived cells of avian embryos

In previous studies, we showed that neural crest (NC)-derived cells from embryonic quail dorsal root ganglia (DRG) and peripheral nerve (PN), which do not normally give rise to melanocytes, become committed to melanogenesis following treatment in culture with the phorbol ester drug 12-O-tetradecanoyl phorbol-13-acetate (TPA). These and other observations support the notion that melanocytes and Schwann cells are derived from a common bipotent intermediate in the neural crest lineage--the melanocyte/Schwann cell progenitor. In this study, we test the possibility that peptide growth factors found in the embryonic environment might act similarly to TPA to influence the fates of these cells. DRG and PN explants were cultured in medium supplemented with a variety of growth factors, and then the cultures were examined for the presence of pigment cells. We found that basic fibroblast growth factor (bFGF), but not various other growth factors, induced pigmentation in about 20% of these cultures. When low concentrations of TPA were included in the culture medium, bFGF augmented the TPA-induced pigmentation, significantly increasing the proportion of pigmented cultures. These effects of bFGF were age-dependent, and could be blocked by addition of a bFGF-neutralizing antibody to the culture medium. In contrast to these stimulatory effects of bFGF, transforming growth factor-beta 1 (TGF-beta 1) was found to inhibit the TPA- or bFGF-induced pigmentation of DRG cultures. These data suggest, therefore, that at least some NC-derived cells are responsive to bFGF and TGF-beta 1, and that these growth factors may play an important role in the control of NC cell fate.     

Adenosine 3',5'-cyclic monophosphate modulates the mitogenic responses of murine B lymphocytes

Adenosine 3',5'-cyclic monophosphate (cAMP) acts to inhibit a number of lymphocyte activities. The extent of this inhibition was tested by evaluating the effects of two cAMP-raising agents on B cell S phase entry induced by several different mitogenic regimens. It was found that both dibutyryl cAMP (dbcAMP) and isobutylmethylxanthine (IBMX) enhanced S phase entry induced by some regimens but inhibited S phase entry induced by others. The observed enhancing activity stands in contrast to the general notion of cAMP as being a "negative regulator," and it confirms that the observed inhibiting activity does not simply reflect cytotoxicity. Mitogenic regimens that appear to mimic each other, such as F(ab')2 fragments of goat anti-mouse immunoglobulin and the combination of a calcium ionophore and a phorbol ester, were distinguished by their responses to the addition of the two cAMP-raising agents. B cell responses were enhanced or inhibited even when dbcAMP was added 18-24 hr after the establishment of cultures. Cyclic AMP may regulate in a complex fashion S phase entry in cells of the immune system.     

Melanocyte stimulating hormone induces the differentiation of mouse epidermal melanocytes in serum-free culture.

In serum-free primary culture of dissociated mouse epidermal cells, alpha-melanocyte stimulating hormone (alpha-MSH) and dibutyryl cyclic AMP (DBcAMP) induced the differentiation of melanocytes. Moreover, the proliferation of melanocytes was also induced in the dishes cultured with DBcAMP, but not with alpha-MSH. In order to clarify the role of keratinocytes in melanocyte proliferation and differentiation, pure cultures of keratinocytes were established in serum-free medium. Subconfluent primary keratinocytes were trypsinized and seeded into pure primary melanoblasts cultured with serum-free medium that did not contain alpha-MSH or DBcAMP. Melanoblasts were cultured with alpha-MSH or DBcAMP in the presence or absence of keratinocytes. alpha-MSH failed to induce melanocyte differentiation in the absence of keratinocytes. DBcAMP failed to induce melanocyte proliferation in the absence of keratinocytes, although it induced melanocyte differentiation even in the absence of keratinocytes. These results suggest that keratinocyte-derived factors are required not only for the induction of melanocyte differentiation by alpha-MSH but also for the induction of melanocyte proliferation by DBcAMP.