Transforming growth factor-beta1 inhibits basal melanogenesis in B16/F10 mouse melanoma cells by increasing the rate of degradation of tyrosinase and tyrosinase-related protein-1.

Current evidence suggests that melanogenesis is controlled by epidermal paracrine modulators. We have analyzed the effects of transforming growth factor-β1 (TGF-β1) on the basal melanogenic activities of B16/F10 mouse melanoma cells. TGF-β1 treatment (48 h) elicited a concentration-dependent decrease in basal tyrosine hydroxylase and 3,4-dihydroxyphenylalanine (Dopa) oxidase activities, to less than 30% of the control values but had no effect on dopachrome tautomerase activity (TRP-2). The inhibition affected to similar extents the Dopa oxidase activity associated to tyrosinase-related protein-1 (TRP-1) and tyrosinase. This inhibition was noticeable between 1 and 3 h after the addition of the cytokine, and maximal after 6 h of treatment. The decrease in the enzymatic activity was paralleled by a decrease in the abundance of the TRP-1 and tyrosinase proteins. TGF-β1 mediated this effect by increasing the rate of degradation of tyrosinase and TRP-1. Conversely, after 48 h of treatment, the expression of the tyrosinase gene decreased only slightly, while TRP-1 and TRP-2 gene expression was not affected. An increased rate of proteolytic degradation of TRP-1 and tyrosinase seems the main mechanism accounting for the inhibitory effect of TGF-β1 on the melanogenic activity of B16/F10 cells.

Current evidence suggests that melanogenesis is controlled by epidermal paracrine modulators. We have analyzed the effects of transforming growth factor-␤1 (TGF-␤1) on the basal melanogenic activities of B16/F10 mouse melanoma cells. TGF-␤1 treatment (48 h) elicited a concentration-dependent decrease in basal tyrosine hydroxylase and 3,4-dihydroxyphenylalanine (Dopa) oxidase activities, to less than 30% of the control values but had no effect on dopachrome tautomerase activity (TRP-2). The inhibition affected to similar extents the Dopa oxidase activity associated to tyrosinase-related protein-1 (TRP-1) and tyrosinase. This inhibition was noticeable between 1 and 3 h after the addition of the cytokine, and maximal after 6 h of treatment. The decrease in the enzymatic activity was paralleled by a decrease in the abundance of the TRP-1 and tyrosinase proteins. TGF-␤1 mediated this effect by increasing the rate of degradation of tyrosinase and TRP-1. Conversely, after 48 h of treatment, the expression of the tyrosinase gene decreased only slightly, while TRP-1 and TRP-2 gene expression was not affected. An increased rate of proteolytic degradation of TRP-1 and tyrosinase seems the main mechanism accounting for the inhibitory effect of TGF-␤1 on the melanogenic activity of B16/F10 cells.
Due to the high reactivity of the melanogenic intermediates, melanogenesis is confined in specialized organelles called melanosomes. Epidermal melanization involves not only the synthesis of the pigment but also the transfer of melanized melanosomes to the keratinocytes surrounding the dendritic processes of melanocytes. Thus, melanocytes are in intimate physical and functional contact with other epidermal cells, which could contribute to the control of the melanogenic status of melanocytes through the secretion of several cytokines and paracrine factors. In keeping with this view, the skin contains a variety of chemical messengers with the potential to affect the proliferation and differentiation of melanocytes. These include several cytokines (13) and proopiomelanocortin-derived peptides (14 -17). Interestingly, the production of these factors by epidermal cells is modulated by environmental conditions, including contact irritants, epidermal wounds, and UV light. It has been shown that some interleukins normally present in the skin are able to inhibit melanization and stimulate proliferation of mouse melanoma cells (18). IL-1␣, IL-6, and TNF-␣ have also been shown to inhibit both melanization and proliferation in normal human melanocytes (19). The mechanisms involved in the inhibition of melanin synthesis by these chemical agents are still unknown.
Here we report a study on TGF-␤1 inhibition of melanogenesis, using B16/F10 mouse melanoma cells as a model. Our results show that this cytokine inhibits in a concentration-dependent way the activities of tyrosinase and TRP-1, with no significant effect on TRP-2 activity. At the concentrations tested, TGF-␤1 was not cytotoxic and the possibility of induction of an enzymatic inhibitor was ruled out. The inhibition was instead explained by a decreased half-life of the melanogenic proteins, and for tyrosinase, by a slight but significant decrease in mRNA levels. These results provide the first description, at the molecular level, of the hypopigmenting effect of TGF-␤1. Moreover, they might provide a framework for the study of more complex regulatory events, like the relationship between hyperpigmenting stimuli, such as MSH and related peptides, and hypopigmenting signals such as TGF-␤1 and other cytokines.

MATERIALS AND METHODS
Cell Culture-B16/F10 mouse melanoma cells were originally a kind gift from Dr. V. Hearing (National Institutes of Health, Bethesda, MD). Cells were cultured in MEM with 10% fetal calf serum and 1% penicillin/streptomycin. When necessary, cytokines diluted in MEM were added to the culture medium 24 h after seeding. Cells were grown to semiconfluence, harvested with 0.1 mg/ml trypsin and 0.2 mg/ml EDTA in Hank's balanced salt solution, and counted before detergent solubilization.
Enzyme Activity Determinations-Cells were washed with Hank's balanced salt solution and solubilized in 1% Nonidet P-40, 0.1 mM phenylmethylsulfonyl fluoride, 10 mM phosphate buffer, pH 6.8. Tyrosine hydroxylase activity was measured by the method of Pomerantz (21) modified as described elsewhere (22). Dopa oxidase activity determinations were performed according to Winder and Harris (23), with minor modifications. DCT activity was quantitated by high performance liquid chromatography as described previously (7). All activity units were defined as the amount of enzyme that catalyzes the transformation of 1 mol of substrate/min.
Electrophoretic Procedures-Analytical SDS-PAGE was performed as described by Laemmli (24) in 12% acrylamide gels. Samples were mixed with sample buffer yielding a final SDS concentration of 3%, but omitting 2-mercaptoethanol and heating, in order to preserve enzymatic activity. Electrophoresis was run at 4°C at a constant current of 25 mA in a Mini-PROTEAN cuvette (Bio-Rad). After electrophoresis, the gels were equilibrated in 50 mM sodium phosphate buffer, pH 6.0, for 15 min and incubated for 15 min at 37°C in a phosphate-buffered solution of 2 mM L-Dopa and 4 mM MBTH (25).
Other Procedures-Protein was determined using the BCA kit (Pierce) and bovine serum albumin as standard. Quantitation of gels, filters, and autoradiograms was performed by laser scanning densitometry, using an Ultroscan XL densitometer from Pharmacia Biotech Inc., and by digital image processing performed with an IMCO 10 image computer from Kontron (Zurich, Switzerland), with comparable results. Cell viability was measured by the MTT assay (34) and by trypan blue dye exclusion.

Effect of Epidermal Cytokines on the Melanogenic
Activities of B16 Cells-The effect of several cytokines normally present in the epidermis on the melanogenic activities of B16 melanoma cells was tested. Treatment of cells with ␣-MSH markedly increased the Dopa oxidase and tyrosine hydroxylase activities (Fig. 1). This treatment was included as a positive control for cell responsiveness. Conversely, IL-1␣ had no detectable effect, whereas TGF-␤1 and TNF-␣ mediated a noticeable decrease of these activities. A similar inhibitory effect of TNF-␣ on the melanogenic activity of normal human melanocytes has been reported recently (19). The effect of the most potent inhibitor identified, TGF-␤1, was further analyzed.
Differential Regulation of the Melanogenic Enzymes by TGF-␤1-After 48 h of culture in the presence of TGF-␤1, both the tyrosine hydroxylase and the Dopa oxidase activities of B16 melanocytes were inhibited in a concentration-dependent way, but DCT activity remained stable (Fig. 2). The effect was apparently mediated by high affinity receptors, since the inhibition was evident at concentrations as low as 10 Ϫ12 M and 50% inhibitions were observed at 10 Ϫ11 M. The cytokine inhibited the Dopa oxidase (and probably the tyrosine hydroxylase) activity of the two tyrosinase isoenzymes, tyrosinase and TRP-1, as shown by electrophoretic separation followed by activity stain (25, 35) (Fig. 3).
The doses of cytokine used were not significantly cytotoxic for B16 cells, as determined by the MTT assay (Fig. 4). The cell number increased during a 72-h incubation period, even at high concentrations of TGF-␤1. However, TGF-␤1 appeared to have a small cytostatic effect on B16 cells after 72 h of treatment. To avoid possible artifacts arising from this effect, all further experiments were performed at shorter incubation times, never exceeding 48 h.
The Inhibition of Tyrosinase Activities Is Not Related to the Induction of a Melanogenic Inhibitor-The possible induction by TGF-␤1 of a tyrosinase inhibitor was ruled out based on the following observations. First, the activity detected was linear with the amount of extract in the assay mixture (data not shown), for control and TGF-␤1-treated cells. This constitutes a kinetic criterion for the absence of effectors in an enzymatic preparation (36). Second, the activity of mixtures of extracts from control and TGF-␤1-treated cells was identical to the theoretical sum of activities from both components (data not shown). Moreover, the residual activity at different TGF-␤1 concentrations was very similar as measured spectrophotometrically or by densitometric quantitation of activity-stained SDS-PAGE gels (Fig. 5). Finally, the K m for tyrosine was identical in extracts from control and TGF-␤1-treated cells (Table  I), and the decrease in V max was consistent with the observed decreases in enzyme abundance (see below).
The Effect of TGF-␤1 Is Rapid and Involves Decreased Intracellular Levels of the Melanogenic Proteins-The inhibitory effect of TGF-␤1 (10 Ϫ10 M) was rapid, since decreases in tyrosine hydroxylase and Dopa oxidase activities were detected as soon as 1 h after addition of the cytokine to the culture medium. The kinetics of inhibition were similar for both activities (Fig. 6). Maximal inhibition was reached around 6 h after addition of TGF-␤1, and then the activity levels remained constant at around 20% of control values.
The inhibition was most likely due to decreased intracellular enzyme levels. This was shown by Western blot experiments using specific antibodies directed against the tyrosinase isoenzymes (Fig. 7). Decreased levels of TRP-1 were detected in TGF-␤1-treated cells, as soon as 30 min after addition of the cytokine. Consistent with previous reports (37), the levels of tyrosinase in control cells, as detected with ␣PEP7, appeared lower than those of TRP-1. Therefore, we used chemiluminescent substrate to enhance the signal. Under these conditions, a decrease in the amount of enzyme could be detected. The decrease in the amount of proteins estimated by quantitation of the blots approximately matched the inhibition of the melanogenic activities, further suggesting that inhibition is the result of a lowered enzyme abundance. Conversely, no changes in the intracellular levels of TRP-2 could be detected with the specific antibody ␣PEP8, even after a 48-h treatment at the highest concentration of TGF-␤1 tested throughout this study.
The Decrease in Tyrosinase and TRP-1 Abundance Does Not Result from Reduced Levels of Their mRNA-As shown in Fig.  8, the levels of TRP-2 mRNA were similar in control and TGF-␤1-treated cells (10 Ϫ10 M, 48 h). No significant changes were also observed for TRP-1 mRNA. Conversely, a moderate decrease (30%) in the abundance of the tyrosinase message was detected. However, this decrease was lower than the one observed for the enzymatic activity and for the protein levels (80%, approximately).

TGF-␤1 Decreases the Half-life of the Melanogenic Activities-
The effect of TGF-␤1 on the half-lives of tyrosinase and TRP-1 was then considered. Cells were pretreated with TGF-␤1 (10 Ϫ10 M, 24 h). After addition of cycloheximide (10 g/ml), cells were collected at different times and their residual tyrosine hydroxylase and Dopa oxidase activities were measured (Fig.  9). Control cells yielded half-lives of 3.2 h and 3.8 h for the Dopa oxidase and tyrosine hydroxylase activities, respectively. Upon treatment with the cytokine, the half-life decreased to values near 1.7 h for Dopa oxidase, and 2.2 for tyrosine hydroxylase. Therefore, in both cases the half-life decreased to values around 50% of the controls. DISCUSSION The occurrence in normal skin of a variety of cytokines and proopiomelanocortin-derived peptides is well documented. Keratinocytes synthesize and secrete IL-1 and TNF-␣ (38), IL-6 (39), and TGF-␤1 (40). Melanoma cells secrete IL-1␣ and ␤, IL-6, IL-8, TGF-␣ and -␤ (41,42), and normal human melanocytes are able to synthesize IL-1␣ and -␤ (43). Moreover, epidermal cells are able to synthesize and process of proopiomelanocortin, thus yielding potentially melanogenic peptides (14 -17). These observations suggest that cutaneous melanogenesis might be regulated, at least partially, by paracrine or autocrine factors. However, although it is increasingly evident that some interleukins could act as paracrine inhibitors of melanogenesis, as well as regulators of melanocyte proliferation, the mechanism(s) by which they mediate their inhibitory effect, as well as their interactions and cross-talk with melanogenic stimuli, are largely unknown.
Among the cytokines tested in this study, both TNF-␣ and TGF-␤1 induced a potent inhibition of tyrosine hydroxylation, the rate-limiting step in melanin synthesis. The inhibition mediated by TGF-␤1 was selected for further study, since TGF-␤1 appears to fulfill several important functions in normal and perturbed mammalian skin. TGF-␤1 supports the growth of skin fibroblasts (44) and inhibits the proliferation of keratinocytes (45,46), although it could also act as a survival factor for these cells by blocking their entry in apoptosis (40). Through the control of the expression of integrins and matrix metalloproteases, it may play a major role in tissue remodeling, angiogenesis, and re-epithelization after cutaneous wounds (47)(48)(49)(50). It has been suggested that wound healing might indeed be a major physiological role of this growth factor (reviewed in Ref. 51), and, accordingly, it is strongly expressed in healing wounds (52).
Our data suggest that TGF-␤1 inhibition of melanogenesis is a specific effect, rather than the result of extensive cellular damage or a nonspecific event following the induction of intracellular proteases. Indeed, although TGF-␤1 inhibits the growth of a variety of cells, this effect does not result from cytotoxicity (reviewed in Ref. 51). B16/F10 cells continued to proliferate, under the conditions used in this study, although at a rate somewhat slower than control cells. Similar results have been reported by others (53). Moreover, most of the experi-

TABLE I Kinetic parameters for the tyrosine hydroxylase activity in control and
TGF-␤1-treated B16 melanocytes B16 cells were cultured with or without 10 Ϫ10 M TGF-␤1 during 48 h. The tyrosine hydroxylase activity of the cellular extracts was measured at increasing substrate concentrations and the kinetic constants were calculated from Lineweaver-Burk plots. V max is given in arbitrary units (dpm/h, under the experimental conditions described under "Materials and Methods"). The experiment was repeated twice with similar results. ments performed throughout this study were conducted at TGF-␤1 concentrations and incubation times such that the cytostatic effect of the cytokine was not detectable. On the other hand, the observation that DCT activity is not affected in TGF-␤1-treated melanocytes lends further support to the view that the inhibition of tyrosine hydroxylase and Dopa oxidase activities is a specific process. DCT is much more sensitive than tyrosinase to proteolysis (5), and should be the melanogenic enzyme more potently inhibited by nonspecific proteolysis.
Several possibilities were considered to account for the specific inhibition of the rate-limiting melanogenic enzymes. A variety of tyrosinase inhibitors of different chemical structure have been described (3). Moreover, modulations of the intracellular levels of such inhibitors have been postulated as a mechanism for tyrosinase regulation by extracellular messengers (54). Therefore we tested whether the inhibition mediated by TGF-␤1 could be accounted for by the induction of a tyrosinase inhibitor. This possibility was ruled out based on a variety of kinetic criteria, and on the observation that the degree of inhibition of tyrosinase activity in extracts from TGF-␤1treated cells was the same as measured under conditions allowing for intermolecular interactions (radiometric and spectrophotometric assays), or after SDS-PAGE separation of the extract. A lower activity in the spectrophotometric assay, as opposed to the electrophoretic method, would be indicative of the presence of an inhibitor in the crude extracts. However, comparable results were obtained with the electrophoretic and spectrophotometric assays. Therefore, it is very unlikely that the inhibitory effect of TGF-␤1 might be mediated by induction of a tyrosinase inhibitor.
On the other hand, the decrease of tyrosine hydroxylase and Dopa oxidase activities could not be explained by a decreased rate of transcription of the tyrosinase and TRP-1 genes, or by an increased rate of degradation of the corresponding mRNA. The levels of mRNA for TRP-1 remained unchanged, whereas those of tyrosinase mRNA only displayed 30% decrease. As opposed to this minor effect on mRNA levels, the abundance of tyrosinase and TRP-1 proteins rapidly dropped following TGF-␤1 treatment. This effect was particularly dramatic for TRP-1, which was barely detectable 6 h after challenging the melanocytes with the cytokine. Therefore, under these experimental conditions, the levels of the tyrosinase and TRP-1 proteins did not match those of their corresponding mRNA. Similar observations have been reported by others for melanocytes treated with a variety of agents able to modulate melanogenesis. For instance, Abdel-Malek et al. (55) have shown that normal human melanocytes respond to ␣-MSH by increasing the intracellular protein levels of tyrosinase, TRP-1, and TRP-2 without any noticeable effect on their mRNA, even after prolonged treatments.
It appears, therefore, that the inhibitory effect of TGF-␤1 on tyrosinase and TRP-1 results from decreased levels of the proteins mainly due to post-translational events, and that this post-translational control of the melanogenic enzymes might be a rather general mechanism of control of pigmentation, not only by hypopigmenting agents, but also by hyperpigmenting stimuli. Our data showing decreased half-lives for tyrosinase and TRP-1 suggest that, at least for TGF-␤1, the post-translational control on these enzymes would be exerted through an increased rate of degradation. This lowered enzyme stability might, in turn, be explained by several mechanisms. On one hand, the "de novo" induction of a specific protease might be envisaged, although the kinetics of TGF-␤1 inhibition appear too rapid, at least for TRP-1, to be dependent on the de novo synthesis of a protein. On the other hand, the melanosomes themselves are a rich source of proteases and other hydrolytic enzymes (56). TGF-␤1 might therefore activate a preexisting protease or, alternatively, mediate an unstabilization of the tyrosinase isoenzymes, thus making them sensitive to such an enzyme. Work is under way to test these possibilities.
In summary, we have shown for the first time that TGF-␤1 is a potent inhibitor of melanin synthesis, through post-translational events leading to decreased half-lives and protein levels of tyrosinase and TRP-1. These data might provide a model for the study of the post-translational control of melanin synthesis in mammals. Moreover, they might provide a framework to understand a variety of pigmentary disorders associated to several physiological and pathological conditions, where an imbalance of paracrine and possibly autocrine melanogenic regulators might result in altered proliferation or differentiation of epidermal melanocytes. Finally, it should be noted that the effects of TGF-␤1 are exerted at low concentrations of the cytokine. Since melanoma cells are able to synthesize and secrete TGF-␤1, our results suggest that an autocrine regulatory loop may contribute to the control of melanogenesis, at least in melanocytes. This possibility will be the subject of further studies.