Contribution of melanogenic proteins to the heterogeneous pigmentation of human melanocytes

Cutaneous pigmentation is the outcome of two events: the synthesis of melanin by melanocytes (MC) and the dona-tion of melanin in melanosomes to surrounding ker-atinocytes (Fitzpatrick and Szabo, 1959). In mammalian MC, the rate of melanin synthesis is determined by intra-cellular enzymes, the best characterized of which are tyrosi-nase, and the two tyrosinase-related proteins (TRP)-1 (gp75 in humans; b-protein in mice), and TRP-2 or DOPAchrome tautomerase (DT) (Hearing and Tsukamoto, 1991). These three proteins are expressed by three distinct genes, yet have remarkable sequence homology (Kwon et al., 1987; Vijayasaradhi et al., 1991; Tsukamoto et al., 1992). Tyrosi-nase seems to catalyze three different reactions in the melanogenic pathway (Körner and Pawelek, 1982). The first two reactions result in the conversion of tyrosine to 3,4-dihydroxyphenylalanine (L-DOPA), and in the subse-quent conversion of L-DOPA to dopaquinone. The third reaction occurs at a distal step in the pathway, in which 5,6-dihydroxyindole is oxidized to indole quinone. The function of TRP-1 is still not clearly defined. It is known that a mutation at the b-locus in mice results in a brown, rather than a black, coat color (Jackson, 1988). The possi-bility exists that in humans the lack of expression of gp75 may lead to a form of tyrosinase-positive oculocutaneous albinism (Zhao et al., 1992). The function of DT is to con-vert DOPAchrome preferentially to dihydroxyindole-2-car-boxylic acid (DHICA), rather than to 5,6-dihydroxyindole (Aroca et al., 1990; Pawelek, 1991). The incorporation of high concentrations of DHICA, rather than 5,6-dihydrox-yindole, into the melanin polymer results in the formation of soluble light brown instead of an insoluble dark brown pigment (Orlow et al., 1992). The contribution of tyrosi-nase, TRP-1 and DT to the diversity and heterogeneity of human pigmentation has not been investigated in depth, and is the subject of the study presented here.

An interesting observation is that the extent of pigmen-tation does not only differ from one individual to another, but is also evident in MC derived from the same skin site of an individual donor. This diversity was first reported by Mishima and Widlan (1967), who noted upon ultrastruc-

tural examination of skin biopsies from donors with dif-ferent skin types that MC within the same biopsy differed significantly in the amount of melanin they contained. In vitro, we have consistently observed heterogeneity in melanin content among MC established from a single skin biopsy. We found that the extent of heterogeneity varied from one cell strain to another, depending on the skin type of the donor, and was mostly pronounced in MC derived from donors with skin types III and IV. We present, for the first time, new data showing that this pigmentary hetero-geneity in human skin can be attributed to the differential expression and/or activity of tyrosinase, TRP-1 and DT by different MC. Also, since cutaneous pigmentation is regu-lated by a multitude of hormones, some of which are paracrine or autocrine in nature, we present evidence that the expression and/or activity of the above three proteins is influenced by such factors (Abdel-Malek et al., 1986, 1987, 1988; Fuller et al., 1987; Swope et al., 1991; Wong and Pawelek, 1973).

We postulate that the heterogeneity of human MC may have significant implications for the susceptibility of dif-ferent individuals to the carcinogenic effects of ultraviolet light. The pattern of melanocytic heterogeneity, which seems to be unique to every individual, and ultimately determines melanin content in the epidermis, may deter-mine the extent of ultraviolet light-induced DNA damage and mutations in the skin.

Pure human MC cultures were established from neonatal foreskins obtained from the nursery of the University Hospital following routine circumcision. Each foreskin was processed individually as follows. The ventral fatty tissue was trimmed and the skin incu-bated in 0.25% trypsin solution (Gibco BRL, Grand Island, NY) overnight at 4°C. The epidermis and dermis were then separated, and both were transferred into a centrifuge tube containing MC growth medium. The tube was vigorously vortexed to obtain a single cell suspension, which was then transferred into a 25 cm² culture flask. The MC growth medium used is a modification of that described by Eisinger and Marko (1982) and consisted of Ham’s F-10 medium, 2.0% fetal calf serum, 2.0% newborn calf serum, 1% penicillin-streptomycin (Gibco BRL), and supple-mented with 5 μg/ml insulin, 2 μg/ml transferrin, 2 μg/ml α-toco-pherol, 0.1 mM 3-isobutyl-1-methylxanthine (IBMX), 2.0 ng/ml cholera toxin and 8 nM phorbol 12-myristate 13-acetate (PMA). Unless stated otherwise, all the above tissue culture reagents were obtained from Sigma Chemical Co. (St. Louis, MO).

Pure cultures of MC were separated according to cell density into subpopulations using a discontinuous Percoll density gradient (Percoll solution was purchased from Sigma Chemical Co.). In each gradient, the following Percoll densities were included: 1.090, 1.080, 1.070, 1.065, 1.060, 1.050, and 1.040 g/ml, with the

highest density located at the bottom of the gradient, as described by Resnicoff et al. (1987). Two million MC suspended in 1 ml sterile phosphate buffered saline (PBS) were layered on top of the gradient and centrifuged at 800 g for 45 minutes at room tem-perature. This allowed MC to be separated into subpopulations, which settled at the equivalent Percoll density. Each subpopula-tion was isolated aseptically, washed once with PBS and then grown in culture.

The tyrosine hydroxylase activity of the separated MC subpopu-lations was compared in situ according to a modification of the charcoal absorption method of Pomerantz (Pomerantz, 1964; Abdel-Malek et al., 1987, 1988; Swope et al., 1991). To measure basal tyrosinase activity, MC were seeded into six cluster wells at a density of 0.94×10⁴ cells/cm² (0.09×10⁶ cells/well) and allowed to proliferate for 6 days. To determine the effect of inter-leukin (IL)-1α and tumor necrosis factor (TNF)-α, MC from Per-coll-derived subpopulations were seeded into a 24-cluster well at a density of 0.05×10⁶ cells/well and treated 24 hours after seed-ing with 10⁻⁹ M TNF-α or 50 units/ml IL-1α, and every other day thereafter for a total of 7 days. TNF-α was a generous gift from Genentech, Inc. (South San Francisco, CA) and IL-1α was purchased from Genzyme Corp. (Cambridge, MA). Twenty-four hours before the end of each experiment, the cultures were sup-plemented with medium containing 1 μCi/ml [³H]tyrosine (specific activity 53 Ci/mmol; Dupont-NEN Research Products, Boston, MA). At the end of the experiment, the medium from each well was collected and assayed for tyrosine hydroxylase activity, and the cell number in each well was determined using a Coulter counter (model ZM). The tyrosine hydroxylase activity was determined by quantitating the amount of ³H₂O released into the medium as [³H]tyrosine is converted to L-DOPA. Each exper-imental group consisted of 3 wells and duplicate 1 ml samples from each well were used for the tyrosine hydroxylase assay (6 determinations per group). Tyrosine hydroxylase activity was expressed as cpm/10⁶ cells per 24 hours.

Melanin content was measured using the method of Lee et al. (1972). Melanocytes were harvested, washed once with phosphate buffered saline (PBS), counted, and then the melanin was solubi-lized in 0.2 M NaOH. Melanin content was determined spec-trophotometrically by reading the absorbance at 475 nm against a standard curve of known concentrations of synthetic melanin (Sigma Chemical Co.), and expressed as μg/10⁶ cells.

For electron microscopic processing, cultured MC were fixed with half-strength Karnovsky’s fixative for 30 minutes at room tem-perature, washed and post-fixed with 1% osmium tetroxide con-taining 1.5% potassium ferrocyanide. For histochemical localiza-tion of tyrosinase, fixed cells were incubated in a 0.1% L-DOPA solution for 2 hours, twice at 37°C. Cells were then dehydrated, embedded in Epon 812, sectioned on an RMC 6000 ultramicro-tome, stained with lead citrate and uranyl acetate, and viewed in a JEOL 100X electron microscope.

The expression of the three melanogenic proteins, tyrosinase, TRP-1 and DT, was determined by western blot analysis. At least 1×10⁶ MC were harvested, pelleted, washed with PBS, and lysed with a buffer consisting of 1% Triton X-100 in 10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 2 mM EDTA, and the protease inhibitors, D-isopropyl fluorophosphate (0.75 μl/ml), Transepoxysuccinyl L-leucylamido (4-guanidino)-butane (E-64; 10 μg/ml), aprotinin (10 μg/ml), pepstatin (10 μg/ml), and leupeptin (10 μg/ml), for 20 minutes on ice, and sonicated for 10 seconds. The cell lysate was subsequently centrifuged at 4°C, at 8,000 rpm for 10 minutes, and the supernatant was used for the experiment. Protein content was quantified using BSA as a standard and the DC protein assay kit (detergent compatible; Bio-Rad, Melville, NY).

An equal amount of protein (8-12 μg) from cell lysates was included in each lane of a Bio-Rad mini-gel electrophoresis apparatus and electrophoresed on 7.5% SDS-polyacrylamide gels. All the reagents for gel electrophoresis were obtained from Bio-Rad. The gels were electrically transblotted on a nitrocellulose mem-brane, and the lanes were reacted with the respective antibodies. For tyrosinase, an anti-hamster tyrosinase antiserum, a generous gift from Dr Seymour Pomerantz, was used at a dilution of 1:10,000 to 1:20,000. For TRP-1, PEP-2 antiserum raised against the amino terminus of the synthetic murine TRP-1 was utilized at a dilution of 1:750, and for DT, PEP-8, an antiserum raised against the carboxy terminus of synthetic murine DT, was used at a dilu-tion of 1:1,000. Both antibodies were a kind gift from Dr Vincent Hearing, Cell Biology Laboratory at NIH. After incubation with the above primary antibodies overnight, the membranes were blocked with 5% non-fat dry milk in basic buffer, pH 7.5, con-sisting of 50 mM Tris, 100 mM NaCl, 0.05% Tween-20, 0.01% sodium azide (3 times for a total of 30 minutes), and the lanes were reacted with alkaline phosphatase-conjugated goat anti-rabbit antibody (Southern Biotechnology Associates, Inc., Birmingham, AL) for 2 hours. The resulting bands were visualized by reacting the membranes with BCIP/NBT alkaline phosphatase substrate (Kirkegaard and Perry Labs, Inc., Gaithersburg, MD).

The specificity of the two antibodies PEP-2 and PEP-8 has been determined by Jiménez et al. (1991). The specificity of the anti-hamster melanoma tyrosinase antibody has been reported by Hal-aban and Moellmann (1990). The lack of cross-reactivity of this antibody with TRP-1 or DT was confirmed using tyrosinase-neg-ative albino MC. We found that lysates from these cells immunoreacted normally with PEP-2 and PEP-8. However, the above tyrosinase antibody, recognized a single protein with a mol-ecular mass greater than 80 kDa, and thus higher than that of wild-type tyrosinase (data not shown).

Human MC in culture exhibit heterogeneity with respect to their melanin content (Fig. 1A). This is readily visible upon inspection of cultures by light microscopy. The extent and pattern of melanocytic heterogeneity (i.e. relative ratio of lightly, intermediately and heavily pigmented MC) is a dis-tinctive characteristic that is unique to every cell strain and thus to each individual. By Percoll gradient centrifugation, an individual cell strain can be separated into several dis-tinct subpopulations that differ from each other in their melanin content (Fig. 1A,B,C). Melanocyte strains estab-lished from type III and IV skin are more heterogeneous and consist of more subpopulations than MC strains established from very fair (i.e. type I) or from deeply pigmented (i.e. type VI) skin (Fig. 2A,B). Comparison of the subpopulations isolated after Percoll gradient centrifugation indi-cated that there was a gradual increase in melanin content and tyrosinase activity with increasing density of the gra-dient (Fig. 3, Table 1; and Bustamante et al., 1991). In all experiments, MC derived from the highest density fraction of the gradient represented the most melanotic subpopulation with the highest tyrosinase activity. Conversely, MC derived from the low density fraction of the gradient were less melanotic and had a lower tyrosinase activity, relative to the rest of the subpopulations. Thus, within the same cell strain, a direct correlation exists between tyrosinase activity and melanin content, as previously shown for MC strains established from individuals with different skin types (Hal-aban et al., 1983; Iwata et al., 1990).

To investigate whether or not the heterogeneity of MC within one cell strain represents a hierarchy in the differ-

entiation state of these cells, we examined the least melan-otic subpopulation by electron microscopy for the presence of undifferentiated melanoblasts. Melanoblasts are expected to be devoid of melanosomes and incapable of synthesiz-ing melanin because they lack tyrosinase activity (Holbrook et al., 1988). We could not detect any cells without melanosomes, and upon incubating the MC with L-DOPA, a positive reaction product, indicative of tyrosinase activity, was observed in the trans-Golgi area, in coated vesicles and in melanosomes of all MC examined (Fig. 4). These results suggest that the difference in the extent of pigmentation among the subpopulations is probably not due to a differ-ence in their differentiation state. Rather, MC with differ-ent melanin content may represent subpopulations within a single cell strain. Evidence for this comes from our obser-vation that when the different subpopulations were followed for 2-3 passages in culture, the differences in melanin con-tent and tyrosinase activity persisted (data not shown).

To determine if the difference in melanin content is also due to a difference in the expression of the melanogenic enzymes tyrosinase, TRP-1 and DT, we compared the amounts of these enzymes in different MC strains from dif-ferent skin types and in the least, intermediate and most melanotic subpopulations, derived from the same MC strain. Using western blotting, we found that MC strains established from Caucasian skin expressed less amounts of the above enzymes than MC strains established from African American skin (Fig. 5). Similarly, the expression of all three proteins increased as melanin content increased in subpopulations derived from the same MC strain (Fig. 6). These results clearly show that the difference in pig-mentation among different individuals and among MC from the same donor can be attributed in part to the difference in the expression of tyrosinase, TRP-1 and DT.

Since the ability of MC to synthesize melanin is influenced by hormonal factors, we compared the ability of the MC subpopulations to respond to the epidermal regulatory cytokines TNF-α and IL-1α (Oxholm et al., 1980; Kupper, 1990) (Table 2). As expected, based on previous results from our laboratory, all the subpopulations responded to both IL-1α and TNF-α with a remarkable decrease in their proliferation and tyrosinase activity (Swope et al., 1991). Also, we found that TNF-α and IL-1α inhibited the expression of tyrosinase, TRP-1 and DT (Fig. 7). From this, it can be concluded that exposure of MC to certain epider-mal factors may affect their melanogenic status by altering the expression of the melanogenic proteins, tyrosinase, TRP-1 and DT.

The variation in skin pigmentation among humans is an obvious trait that is under strict genetic control. We found that cultured MC derived from the same skin biopsy are heterogeneous in the amount of melanin they contain. This was observed not only in neonatal foreskin MC, but also in MC derived from different anatomical sites such as fore-skin, forearm, breast and thigh, of children as well as young and middle-aged adults (data not shown). This functional heterogeneity is observed in MC cultured under different culture conditions, and has previously been noted morpho-logically by electron microscopy of skin in vivo (Mishima and Widlan, 1967). The degree and pattern of heterogene-ity seem to be characteristic of each individual MC strain. Heterogeneity is mostly pronounced in MC strains derived from skin types III and IV (Figs 1A, 2). We have suggested that the variation in the degree of melanization of MC within the same cell strain is due to the differential expression and/or activity of the three known melanogenic proteins, tyrosinase, TRP-1 and DT. Also, the differential responsiveness of MC with different degrees of melaniza-tion to paracrine/autocrine factors might contribute to the difference in melanin content by affecting the activity and/or synthesis of the above proteins.

To investigate the above hypothesis, we have used the method of Percoll density gradient centrifugation to sepa-rate heterogenous MC strains into subpopulations. Another method that has been successfully used for the same pur-pose is flow cytometry, i.e. sorting, of MC with different degrees of pigmentation according to their light scattering ability. Deeply pigmented MC cause minimum, while lightly pigmented MC cause maximum, forward light scat-ter (Boissy et al., 1990). The number of the subpopulations derived following Percoll density gradient centrifugation depended on the extent of heterogeneity of the original cell strain (Fig. 2A,B). Each MC subpopulation in itself was for the most part homogeneous in its pigmentation, but differed from the other subpopulations in its melanin content and tyrosinase activity (Fig. 1B,C, Fig. 3, Table 1; and Busta-mante et al., 1991). The least melanotic MC subpopulation, which was derived from the low density fraction of the gra-dient, had the lowest tyrosinase activity, and vice versa. By western blotting, we found that the expression of tyrosi-nase, TRP-1 and DT also correlated directly with the amount of melanin and the Percoll density of each sub-population (Fig. 6, Table 1). The quantity of each of these enzymes was significantly greater in the most melanotic subpopulation with the highest density than in the lighter and less dense subpopulations. These same differences in tyrosinase activity and the expression of tyrosinase, TRP-1 and DT were also found among cell strains derived from Caucasian or African American neonatal foreskins (Fig. 5). This is the first documentation of the role of the three melanogenic proteins in determining the extent of pigmen-tation of human melanocytes.

The direct correlation that we observed between tyrosinase activity and melanin content was demonstrated previously by Halaban et al. (1983), and Iwata et al. (1990). Also our finding that tyrosinase is more abundant in darkly than in lightly pigmented MC is in agreement with the results of Halaban et al. (1983). In our studies we have extended these observations and showed that not only tyrosinase, but also TRP-1 and DT, are more abundant in deeply than in lightly pigmented MC. This correlation has also recently been demonstrated for murine melanoma cell lines that differ in their extent of melanization (Kameyama et al., 1993). Melanocytes with a low melanin content and tyrosi-nase activity were found to process tyrosinase more slowly and degrade it more quickly than MC with a higher melanin content and tyrosinase activity (Halaban et al., 1983). The expression of tyrosinase in normal human MC seems to be regulated at a post-transcriptional step, since Naeyaert et al. (1991) have shown that the tyrosinase-specific mRNA level did not correlate with melanin content or tyrosinase activity. Whether or not the difference in the abundance of the three melanogenic proteins in the MC subpopulations derived from a single cell strain is due to differences in the rates of protein processing and degradation or the level of gene transcription is still to be investigated.

Cellular heterogeneity is characteristic of may cell types, normal and malignant. Heterogeneity may be a manifesta-tion of differentiation hierarchy, as in the case of human breast tumor cell line MCF-7 (Resnicoff et al., 1987). Such a hierarchy is also exemplified in the skin by keratinocytes, which become progressively more differentiated as they migrate from the basal to the granular layer of the epidermis (Fuchs, 1990). Alternatively, heterogeneity may repre-sent multiple clonogenicity; for example, tumors that are composed of different clones, each of which had originated from a single progenitor or stem cell.

To determine whether or not the MC subpopulations within one strain represent different differentiation stages, we examined the least melanotic subpopulation for the presence of undifferentiated melanoblasts. A melanoblast is characterized by the lack of ability to form melanosomes and synthesize melanin due to the inactivity or absence of melanogenic proteins (Holbrook et al., 1988). Using bio-chemical techniques, we found that lightly pigmented cells contain melanin and tyrosinase activity (Fig. 3, Table 1). By electron microscopy, we confirmed the presence of melanosomes and an active tyrosinase enzyme within these MC, as depicted by deposition of DOPA reaction product in the trans-Golgi area, coated vesicles and melanosomes (Fig. 4). These results do not totally negate the possibility that a subpopulation of melanoblasts exists. This popula-tion, however, could not be found by our methods or did not persist in culture. These results strongly suggest that the difference in the melanogenic status of the different sub-populations does not represent a difference in their state of differentiation.

It is known that epidermal MC are influenced by a vast number of hormones. Among these are soluble factors that are released by keratinocytes and act as paracrine regulators, affecting the proliferation and the melanin synthetic rate of MC (Gordon et al., 1989). We have recently described the inhibitory effects of the three epidermal immune/inflammatory cytokines IL-1α, IL-6 and TNF-α (Swope et al., 1991). It has been shown that all three cytokines are synthesized by human MC and, therefore, might act as autocrine/paracrine mediators (Mizutani et al., 1990; Köck et al., 1991; Swope et al., 1993, unpublished data). We have found that neonatal human MC derived from male skin samples (neonatal foreskins), respond in a simi-lar manner to adult human MC from female skin samples (breast skin) to the above three cytokines as well as to a variety of growth factors (data not shown). Using MC derived from male and female skin samples excludes the possibility that heterogeneity is the result of differential responsiveness of MC to steroid sex hormones. To deter-mine the role of autocrine/paracrine factors in the hetero-geneity of MC, we investigated the responsiveness of dif-ferent MC subpopulations to TNF-α and IL-1α. All the subpopulations tested demonstrated diminished prolifera-tion and tyrosinase activity in response to these two cytokines (Table 2). Both cytokines elicited their inhibitory effect on melanogenesis primarily by inhibiting tyrosinase activity, and by decreasing the expression of tyrosinase, TRP-1 and DT (Table 2, Fig. 7). This suggests that the degree of melanogenic heterogeneity of MC is controlled to some extent by their responsiveness to paracrine/autocrine factors.

The significance of MC heterogeneity is unknown. Melanin is presumed to function as a sunscreen, and to neu-tralize oxygen radicals generated by a variety of factors, including ultraviolet light B (UVB) (Menon and Haberman, 1977; Nordlund, 1985). Therefore, melanin may provide protection against UVB-induced DNA damage. On this basis, we speculate that the relative abundance of highly to lightly melanotic MC in the skin might dictate the suscep-tibility of an individual to UV light-induced DNA damage and possibly carcinogenesis.

The authors thank Dr Vincent Hearing for kindly providing antibodies against TRP-1 and DT, and Dr Seymour Pomerantz for his kind gift of anti-tyrosinase antibody. We are also grateful for Mrs Ying L. Boissy for preparing the light micrographs, Dr Estela Medrano for her helpful suggestions, and Ms. Joan Griggs for her excellent secretarial assistance. This work was supported in part by The Dermatology Foundation, Syntex Laboratories Research Grant, Evanston, Illinois; The American Cancer Society Institutional Research Grant (Z.A.M.); and The National Vitiligo Foundation, Tyler, Texas (J.N.).