MSH inhibits growth in a line of amelanotic hamster melanoma cells and induces increases in cyclic AMP levels and tyrosinase activity without inducing melanogenesis

Melanocyte stimulating hormones a, ß and γ (MSH or melanotropins) are neuropeptides with multiple regulatory functions. Along with ACTH and ß-LPH, they are synthesized as part of a large pituitary protein precursor, proopiomelanocortin (POMC) (Eberle, 1988; Hadley, 1988). Today the best-described mechanism of action of MSH is in the regulation of pigmentation, shape and proliferation of rodent melanoma cells (Lerner et al. 1979; Moellmann etal. 1988). Transduction of the melanotropic signal is associated with cell surface receptors (Varga et al. 1974) that are coupled to guanine nucleotide binding proteins (G-proteins) (Gerst & Salomon, 1987), with stimulation of adenylate cyclase (Bitensky etal. 1972), hormone internalization (Varga et al. 1976) into premelanosomes and lysosomes (Lerner et al. 1979) and the nucleus (Moellmann & Lambert, unpublished results), and activation of kinases (DeGraan etal. 1987; Pawelek et al. 1985).

The role of the subcellular apparatus of melanogenesis is to oxidize L-tyrosine into melanin in a precise and efficient way, without destruction of cells by toxic intermediates of the reaction sequence. This process takes place in melanosomes (Seiji, 1967) and is, at least in part, controlled by tyrosinase (monophenol dihydroxyphenylalanine: oxygen oxidoreductase EC 1.14.18.1), which catalyses the hydroxylation of the primary substrate L-tyrosine to L-dopa, and L-dopa to dopa quinone (Lerner et al. 1949). Recent reports have shown that synthesis and function of the melanogenic apparatus, including tyrosinase, can be controlled by L-tyrosine and L-dopa (McEwan & Parsons, 1987; Sato et al. 1987; Slominski et al. 1988), which is in addition to the well-described control by MSH.

Several melanoma models are being used to study regulation of phenotype by MSH (Eberle, 1988; Gerst & Salomon, 1987; Hadley, 1988; Legrose/ al. 1981; Lerner etal. 1979; Moellmann et al. 1988; Pawelek et al. 1985). One of these is the Bomirski Ab amelanotic melanoma, which has the ability to differentiate in cell culture into pigment-producing cells under pressure of environmental stimuli that include the above two substrates in the melanogenic pathway, L-tyrosine and L-dopa (Bomirski & Slominski, 1986; Slominski, 1985; Slominski et al. 1984, 1988). A subline of this tumour, AbCl, established in vitro, has been shown to have MSH receptors that respond to the presence of L-tyrosine by changes in binding capacity and receptor activity (Slominski & Pawelek, 1987; Slominski et al. 1989b).

In this paper we describe a partial regulation of phenotype by MSH in Bomirski AbCl melanoma cells and compare this mechanism with that controlled by L-tyrosine and L-dopa.

The first and only existing subline to date, AbC1, of the Bomirski Ab amelanotic melanoma was established in culture in 1985 from the Bomirski Ab amelanotic melanoma (passage no. 518) through cloning of dispersed tumour cells in soft agar according to Pawelek et al. (1974). The cells were isolated from solid tumour tissue by a non-enzymic method (Slominski, 1983). Monolayer cultures of the AbCl line were grown in Ham’s F-10 medium (GIBCO Labs), containing calf serum (10%) (GIBCO Labs), streptomycin (11 μgml^-1), and penicillin (100units ml^-1). Stock cultures were passaged weekly and supplied with fresh medium three times per week. Swiss ?T? cells served as controls in some experiments.

Rates of cell proliferation in monolayer culture were determined by inoculating 4× 10⁴ cells into 25 cm² flasks in 5 ml Ham’s F-10 medium plus 5 % calf serum, with or without addition of IBMX (isobutylmethyl xanthine, 500μM), dbcAMP (dibutyryl cyclic AMP; lmM) or HPLC-pure β-MSH (lμM) or TPA (lnM, 10nm, 100 nM; 12-O-tetradecanoyl-phorbol 13-acetate, Consolidated Midland Corporation). Thereafter, cultures were supplied with fresh medium daily. The cells were harvested in Tyrode’s balanced salt solution, containing EDTA (1 niM), and counted in a model Z_?1 Coulter Counter (Coulter Electronics Inc., Florida) in which the flow of single cells was controlled by the aperture.

To measure anchorage-independent clonal growth, the cells (500 cells/tube) were inoculated into polystyrene round-bottom tubes (16mm × 125 mm, Becton Dickinson) containing Ham’s F-10 medium plus 15% calf serum, 0·12% Noble Special Agar (GIBCO Labs), and different concentrations of /3-MSH (0·1-1000nm) or IBMX (500μM). After 3 weeks, macroscopically (by eye) visible clones were counted.

Melanoma cells were inoculated into 25 cm² flasks (5× 10^s cells/ flask) in 5 ml Ham’s F-10 medium plus 10% serum, with or without addition of IBMX, dbcAMP, β-MSH or TPA. Thereafter, fresh medium was supplied daily. After 4 days of culture, the cells were harvested and pelleted by centrifugation. The pellets were then lysed in sodium phosphate buffer (0·1 M, pH6·8) containing Triton X-100 (0·5%). Cell extracts were centrifuged at 30 000g for 3 0 min at 4°C and the supernatants used for the in vitro assays. Both the tyrosine hydroxylase (conversion of L-tyrosine to L-dopa) and dopa oxidase (conversion of L-dopa to dopa quinone) activities of tyrosinase were measured as described (Pawelek, 1978; Pomerantz, 1964, 1969; Slominski etal. 1988). Briefly, tyrosine hydroxylase activity was assayed at 37°C by the Pomerantz method by measuring ³H₂O released from L-[³H]tyrosine during the enzymic tyrosine hydroxylation, with 0·5μCi of ³H-labelled 3,5-L-tyrosine (55·7 Ci mmol^-’, New England Nuclear Corporation) and 0·1 mM-L-tyrosine (Sigma) as substrates, and 0·1mM-L-dopa (Hoffmann-LaRoche) as cofactor in 0·l M-phosphate buffer, pH6·8. The total volume of reaction mixture was 35 μl. The reaction was stopped by the addition of 1 ml of activated charcoal (10% in 0·1 M-citric acid), and the tubes were centrifuged. The supernatants were passed through Dowex 50 W resin (AG 5OW-X8, 200-400 mesh hydrogen form, Bio-Rad) columns, the eluents containing ³H₂O were collected in scintillation vials and mixed with Optifluor (Packard). Radioactivity was measured in a Beckmann LS 7000 counter. Tyrosine hydroxylase activity was expressed as nmol tyrosine hydroxylated mg⁻¹ protein h⁻¹. Dopa-oxidase activity was spectro-photometrically measured in a total volume of 1ml, in 1-cm light path cuvettes, at 475 nm at room temperature, with 1 mM-L-dopa as substrate in 0·1 M-phosphate buffer, pH6·8 (Bomirski et al. 1987). Increase of absorbancy was measured in a Varian DMS8O spectrophotometer. To calculate the dopaoxidase activity the molar absorbancy coefficient of dopachrome (3600) was used. Dopa-oxidase activity was expressed as nmol dopachrome produced mg⁻¹ protein h⁻¹. Protein content was determined by the method of Bradford (1976).

Tyrosinase activity in living cells was measured by the method described by Fuller (1981). The cells were cultured for 4 days in Ham’s F-10 medium plus 10% calf serum, with or without addition of IBMX (500μM) or β-MSH (1 μM)· The cells were harvested and then inoculated into 25 cm² flasks at 5×10⁵ cells/flask in ?ml medium consisting of Ham’s F-10 medium plus 10% calf serum and 0·5μCiml−1 of L-[³H]tyrosine, with or without addition of IBMX (500μM) or β-MSH (1 μM)· The flasks were capped tightly in order to prevent potential escape of ³H₂O vapour and incubated at 37°C for 24h. Tyrosinehydroxylase activity was measured in terms of tritium release as ³H₂O per 1 ml of medium. ³H₂O was separated from [³H]ty?o-sine as described for the in vitro tyrosinase assay above. Swiss ?T? cells, adapted to grow continuously in Ham’s F-10 medium, served as a control for non-specific tyrosine oxidation.

For ultrastructural localization of the dopa-oxidase activity of tyrosinase, cultures were fixed in situ in a mixture of glutaraldehyde (2·5 %) plus formaldehyde (2%) in 0·1 M-sodium cacodylate buffer, pH 7·2, at 4°C. The cells were then scraped off the culture surface, pelleted and processed as described (Slominski etal. 1988).

Melanoma cells were inoculated into Castor 12-well cluster tissue culture plates at 0·5× 10⁶cells/well in 1ml medium consisting of Ham’s F-10 medium plus 10 % calf serum, without any further additions in order to maintain the amelanotic phenotype, or with the addition of L-tyrosine (200 μM) to change the phenotype to melanotic (Slominski etal. 1988). After 48 h the media were changed to 0·5ml Ham’s F-10 medium plus 10% calf serum, and after 4h the cells were exposed for 10min to L-tyrosine (200μM), L-dopa (50μM) or ß- MSH (0·1 μM). Then the medium was aspirated, the reaction was stopped by addition of 0-5 ml cold 5 % trichloroacetic acid (TCA), and the plates were held at 4°C for 30min. The acidified solutions were collected, the wells were washed once again with 1 ml of TCA and the solutions pooled and extracted five times with 6vols of ether and then lyophilized. After redissolving the lyophilate in sodium acetate buffer, cyclic AMP or cyclic GMP were measured according to Steiner et al. (1972) with the aid of Rianen cyclic AMP and cyclic GMP kits from New England Nuclear Corporation.

Melanoma and ?T? control cells were inoculated into Castor 12-well cluster tissue-culture plates at 0·5×10⁶ cells/well in 1 ml of medium consisting of Ham’s F-10 medium plus 10% calf serum, without any additions in order to maintain the amelanotic phenotype, or with the addition of L-tyrosine (200 μM) to change the phenotype to melanotic. After 48 h the cells were labelled for 24 h with [myo-2-³H]inositol (20 Cimmol⁻¹, Amer-sham) at a concentration of l5μCiml⁻¹ of medium, consisting of Ham’s F-10 medium plus 10% calf serum plus or minus L-tyrosine (200μM). After washing three times with Ham’s F-10 serum-free medium plus LiCl (10mM) and a further 30 min of incubation in the washing medium, the cells were exposed to L-tyro?ine (200μM), L-dopa (50 μM) or β-MSH (0μlμM) for 30s, 2 min or 5 min. The reaction was stopped by addition of cold 10% perchloric acid (PCA) plus 1 mM-diethylenetriamine-penta-acetic acid (DTPA) (Sigma) and ?mM-EDTA. After neutralization with KOH and centrifugation at 4°C the tubes were placed in ice, and the supernatants were aspirated carefully, in order not to remove any salt. The supernatants were diluted up to 5 ml with distilled deionized water and applied to AG 1-X8 Dowex columns (200-400 mesh, formate form, BioRad). InsP3 was collected according to Berridge et al. (1983). Inositol-l,4,5-trisphosphate, labelled at D-inositol-l[₃H](N) (20 Ci mmol^-1, New England Nuclear Corporation) was used as a recognition standard in the collection of the InsP₃ fractions.

Exposure of AbCl melanoma cells to β-MSH and agents that increase intracellular levels of cyclic AMP, i.e. dbcAMP or IBMX, caused: (1) an increase of dendrite formation (Fig. 1); (2) an inhibition of cell growth in monolayer (Fig. 2); (3) an inhibition of the ability of the cells to form macroscopically visible clones in soft agar (Table 1); and (4) a stimulation of tyrosinase activity as measured in cell extracts (Table 2). Tyrosinase activity in vivo was not detectable and not inducible by MSH or IBMX (data not shown), nor did MSH, dbcAMP or IBMX induce the biosynthesis of melanin (Figs 3-5).

The stimulation of tyrosinase activity by MSH (1 μM) or IBMX (500μM), as measured in cell extracts, was reflected in strong dopa reactions at the level of ultrastructure (Fig. 4). The copious reaction product was, however, restricted to trans-Golgi cisternae and associated tubules and vesicles (írans-Golgi reticulum, or TGR). Newly formed organelles, resembling granular melanosomes stage II, were dopa-negative (Fig. 5). The cells contained no melanin. For comparison, in cells induced to melanize by L-tyrosine, the dopa reaction was minimal in the TGR and localized mostly within highly melanotic granular melanosomes (Slominski et al. 1988).

In order to test whether signal transduction from MSH, L-tyrosine and L-dopa occurred along similar or different pathways, we tested levels of three second messengers linked to G-proteins, cyclic AMP, cyclic GMP and InsP₃ (Berridge etal. 1983). As expected, MSH raised cyclic AMP levels in a time-dependent manner (Fig. 6), with a likely peak between 10 and 30 min. The measured peak represented a quasi-10-fold elevation from basal level. A 10-niin exposure was chosen for further comparative studies of the effects of L-tyrosine and L-dopa on cyclic AMP levels. The basal levels of cyclic AMP in amelanotic and melanotic AbCl melanoma cells were similar (Table 3). L-tyrosine stimulated cyclic AMP production by 44%, but only when added to amelanotic cultures. This stimulation was statistically significant, 0·005 >P> 0·0025 according to Student’s t- test. L-dopa (50 μM) did not have any effect on cyclic AMP levels after 10min (Table 3), or after 1, 30 or 60min (data not shown).

MSH, L-tyrosine and L-dopa had no effect on the levels of cyclic GMP, measured over a time span of 10 min (Table 4), or InsP₃, measured over time spans of 30s, 2 min (Table 5) and 5 min (data not shown).

We also tested for an effect of TPA on melanogenesis and cell proliferation in AbCl cells over a period of 4 days (data not shown), because TPA is mitogenic for normal avian, murine and human melanocytes (Halaban, 1988) or can inhibit growth and induce cell differentiation of human melanoma cells (Halaban et al. 1986; Huberman et al. 1979; Loms Ziegler-Heitbrock et al. 1985). TPA at concentrations of 1, 10 and 100nm induced neither melanin synthesis (cell pellets were white) nor tyrosinase activity as measured in cell extracts. TPA at these concentrations did not have any effect on cell growth.

In this report we have shown that in Bomirski AbCl melanoma cells MSH increases intracellular cyclic AMP, and that the hormone and other agents that raise intracellular levels of cyclic AMP inhibit cell growth, induce elongation of dendrites, and substantially increase tyrosinase activity and formation of organelles resembling granular premelanosomes stage II. Neither MSH, dbcAMP nor IBMX induces melanogenesis. MSH, therefore, induced only partial cell differentiation, most probably by activation of a cyclic AMP-dependent regulatory mechanism.

Most previous reports on growth inhibition and stimulation of melanogenesis in rodent or human melanoma cells by MSH or other agents that raised the levels of cyclic AMP were based on studies with melanotic melanomas of different degrees of pigmentation (Giuffre etal. 1988; Legros etal. 1981; Lerner et al. 1979; Moellmann et al. 1988). Increases in tyrosinase activity in these systems were associated with stimulation of melanogenesis. In the past, the close association of tyrosinase activity and melanogenesis has been taken for granted, so much so that as a rule cells that did not respond to MSH or its second messenger by grossly visible increases in pigmentation have probably not been tested for increases in tyrosinase activity. As a consequence, growth inhibition by MSH in actively melaniz-ing cells is easily interpreted as resulting from the cytotoxicity of intermediates of melanogenesis, especially since there are instances in which MSH has been shown to be mitogenic to pigment cells when the toxic intermediates of melanogenesis are washed away (Halaban & Lerner, 1977). In Bomirski AbCl melanoma cells, on the other hand, activation of the cyclic AMP system is effectively increasing tyrosinase activity without inducing melanogenesis. The inhibition of cell growth by MSH in this system is, therefore, independent of melanogenesis.

Previously we had shown that in the Bomirski melanoma system L-tyrosine and L-dopa increased tyrosinase activity (an increase that was inhibited by cycloheximide) and induced the synthesis of melanosomes and melanin (Slominski et al. 1988). In contrast to MSH, L-tyrosine, and to a lesser extent L-dopa, permitted the translocation of tyrosinase (and acid phosphatase) from the TGR to newly synthesized (pre)melanosomes. In cells treated with IBMX or MSH tyrosinase activity remained restricted to the TGR, thereby explaining the failure of living cells to melanize or to have demonstrable tyrosinase activity in vivo. Melanin was not synthesized, even though new organelles resembling granular premelanosomes stage II were formed, especially with IBMX. Such granules are rare in untreated cells of the AbCl line but are induced by L-tyrosine before the onset of L-tyrosine-induced melanization (Slominski etal. 1989a). We recently encountered another amelanotic cell type, i.e. melanocytes cultured from newborn hybrid chinchilla/ albino mice that also carried the pink-eye mutation (c^chp/cp). These amelanotic pigment cells had measurable tyrosinase activity in vitro but not in vivo. The dopa reaction in aldehyde-fixed cells was restricted to the TGR despite the presence of large numbers of premelanosomes (stage II) (Halaban etal. 1988). The latter remained amelanotic even in the presence of dopa. A similar site restriction for tyrosinase has been reported recently for chinchilla hair-bulb melanocytes (Imokawa et al. 1988).

Failure of tyrosinase-rich cells to melanize may be rationalized for those cells that do not synthesize premelanosomes, such as Ab amelanotic tumour cells inside an animal host. The tyrosinase has no final organelle in which to be active (Bomirski etal. 1988). Likewise, during the induction of tyrosinase increases by L-dopa, when melanin production lags behind that induced by L-tyrosine, the lighter colour of the cells can be rationalized on the grounds that L-dopa induces fewer melanosomes than does L-tyrosine (Slominski et al. 1988). The lack of translocation of tyrosinase in AbCl cells, which synthesize both tyrosinase and premelanosomes, cannot be readily explained. The fault may lie with premelano-somes but more likely with the tyrosinase molecule itself and the transporting vesicles that originate from the TGR. In all likelihood, an intrinsic recognition message that normally signals the fusion between premelanosomes and the tyrosinase-containing vesicles is absent or non-functional, and the enzyme accumulates in the cisternae, tubules and vesicles of the TGR.

Measurements of cyclic AMP and cyclic GMP show that L-tyrosine and L-dopa have little or no effect on levels of free intracellular cyclic nucleotides in the model under study. Therefore, the ready induction of the melanotic phenotype by precursors of melanin may follow a cyclicnucleotide independent pathway. Participation of InsP₃-dependent and TPA-inducible pathways has also been almost ruled out by our experiments. With respect to MSH we have ruled out the possibility of hormonal signal transduction through two other G-protein linked systems generating second messengers, i.e. cyclic GMP and InsP₃. The possibility that the MSH receptor is coupled to phospholipase C, in addition to an adenylate cyclase, has been theoretically considered by others (Eberle, 1988; Hadley, 1988; Lucas et al. 1987). The lack of an MSH effect on InsP₃ turnover in hamster malignant melanocytes (presented above) and rat adrenocortical cells (Hyatt etal. 1986), shows that in at least two mammalian systems a direct activation of phospholipase C by MSH does not occur.

In conclusion, then, we have demonstrated that in Bomirski AbCl melanoma cells MSH induces only partial cell differentiation in association with a peak increase in cyclic AMP levels to about 1000% of basal level. Induction of melanization by L-tyrosine and/or L-dopa appears to follow pathways unrelated to elevations of cyclic AMP, cyclic GMP or InsP₃.

This research was supported by US PHS grants 5 R01 CAO4679 to. A. B. Lerner,and 5T32AR07016 to Richard L. Edelson and Aaron B. Lerner, principal coinvestigators. We thank Dr Lerner for his support and thank Jack Schreiber for assistance in the preparation of the manuscript.