Melanocyte Differentiation Marker gp75, the Brown Locus Protein, Can Be Regulated Independently of Tyrosinase and Pigmentation

Melanocyte Differentiation Marker gp75, the Brown Locus Protein, Can Be Regulated Independently of Tyrosinase and Pigmentation Setaluri Vijayasaradhi, Peter M. Doskoch, Jedd Wolchok, and Alan N. Hougliton Immunology Program and Department of Medicine, Memorial Sloan-Kctteritig Cancer Center, New York, New York, U.S.A.

Human melanoma arises from epidermal melanocytes and displays remarkable phenotypic heterogeneity. This heterogeneity in part reflects the ability of melanoma cells to undergo differentiation along a pathway parallel to differentiation of normal melanocytes. Tyrosinase, encoded by the albino (c), and the tyrosinase-related protein-1 or gp75, encoded by the brown (b) locus, are two of the best-characterized markers for melanocyte differentiation. Both molecules are glycoproteins expressed in melanosomes, the site of pigment synthesis. W^e studied the regulation of these proteins in human melanoma cells induced by the polar-planar compound hexamethylene bisacetamide (HMBA). In well-differentiated melanoma cell lines, HMBA induced dendritic morphology and specifically regulated the expression of melanosomal glycoproteins (but not a panel of other molecules expressed by melanoma cells). HMBA specifically down-regulated gp75 expression by rapidly decreasing the steady-state level of gp75 mRNA and

M

elanocytcs are derived by difFerentiation of precursor melanoblasts that originate and migrate from neural crest [1]. Progressive differentiation of melanoblasts in the skin during embryonic and neonatal development produces pigmented, dendritic melanocytes ofthe adult epidermis. Three distinct stages in melanocyte differentiation—early, intermediate, and late—have been proposed based on the expression of a set of cell surface and intracellular markers by cultured fetal, newborn, and adult human melanocytes and melanoma cells [2]. The most striking characteristic of mature melanocytes is the presence ofthe pigment melanin witliin .specialized organelles, melanosomes. The appearance of a pigmented phenotype is accompanied by the expression of a set of melanosomal proteins and other melanocyte-specific markers [3]. The albino (c) locus protein, tyrosinase, and the brown (b) locus protein, tyrosinase-related protein-l/gp75, are two of the bestcharacterized molecules expressed by mature human melanocytes [4_8]. Both tyrosinase and gp75 are 75-kD integral membrane glycoproteins localized to the melanosome [9,10]. Tyrosinase is the

gp75 synthesis. HMBA was able to down-regulate gp75 expression even in the presence of cholera toxin, w^hich when added alone induced a two- to threefold increase in gp75 expression. In contrast to uniform dovt^n-regulation of gp75 expression, HMBA could either up-regulate or down-regulate tyrosinase expression and pigmentation. Based on the differential regulation of gp75 and tyrosinase, melanoma cells could be classified into two groups. In one group, gp75 expression vi^as coordinately regulated with tyrosinase activity and pigmentation. In the other group, gp75 expression and tyrosinase activity and pigmentation w^ere dissociated (v»rith pigmentation coupling to tyrosinase activity, not to gp75 expression). These results show that in mature melanocytic cells, regulation of gp75 expression follows a pattem that can be independent of regulation of tyrosinase and pigmentation. Key words: tnelauosomal proteinslhexatnethylene bisacetatnidelcyclic AMP. J Invest Dermatol 105:113-119, 1995

critical enzyme required for melanin synthe.si.s [6,11]. The bivii'ii locus in mice influences the type of melanin pigment (black versus brown) and produces different shades of coat color [12—14]. Studies with pharmacologic agents that induce phenotypic changes in melanoma cells in i'itro have provided a basis for understanding melanocyte and melanoma differentiation [3]. These studies have shown that human melanoma cells have an ability to differentiate along a pathway that parallels normal melanocyte differentiation. Accordingly, three classes of melanoma cell lines can be identified corresponding to the early, intermediate, and mature melanocytes [2]. These studies have also sbown that difFerentiation is accompanied by coordinated and sequential changes in phenotypic traits. When individual clones of early stage (gp75 and tyrosinase negative) melanoma cells are induced to differentiate by phorbol ester, an activator of protein kinase C (PKC), and/or by cholera toxin, which elevates cellular cyclic adenosine monophosphate (cAMP) levels, pigment synthesis and gp75 are up-regulated and cells acquire dendritic morphology [2]. Very little is known about the regulation of tyrosinase and gp75 in differentiated melanocytes. For example, coincident with decreased growth of mature melanoma cells, down-regulation of gp75 has been observed |2]. To understand the regulation of melanocyte differentiation products, we studied the effect of hexamethylene bisacetamide (HMBA) on the expression of gp75 and tyrosinase in pigmented human melanoma cells. HMBA is a

Manuscript received September 28, 1994; final revision received March 24, 1995; accepted for publication April 3, 1995. Reprint requests to: Setaluri Vijayasaradhi, The Rockefeller University, Box 178, 1230 York Avenue, New York, NY 10021. Abbreviation: HMBA, hexamethylene bisacetamide.

0022-202X/95/S09.50 • SSDI0022-202X(95)00203-W • Copyright C) 1995 by The Society for Investigative Dermatology, Inc. 113

114

VIJAYASARADHI L"r /!L

polar-planar compound that has been shown to induce differendation of a number of murine and human tumor cells iu vitro [15—19]. hi Cloudman S91 murine melanoina cells, HMBA has been shown to inhibit tyrosinase activity and melanogenesis induced by melanocyte-stimulating hormone and isobutylmetbylxanthine [18]. In this report, we show that HMBA specifically down-regulates gp75 in human melanoma cells and that regulation of gp75 expression can b e independent of tyrosinase and pigmentation. MATERIALS AND METHODS Cell Culture Human melanoma cells SK-MEL-19, SK-MEL-23 (clone 22), and all other melanoma cells used were maintained as descrihed earlier [8]. For experiments to study the effect of HMBA on growth, 2 X 10'' cells/well were plated in 24-well plates (Costar) in 1 ml of growth medium. After allowing the cells to attach for 24 h, we replaced the medium with fresh growth medium (control) or medium containing 1, 3, or 5 mM HMBA. Cells were fed every .second day with medium (including FIM13A). Cell growth was monitored hy counting viable cells using the trypan blue exclusion method. Cells from six wells for each treatment on days 1, 3, 5, and 6 were detached using 0.5 ml trypsin/ethylenediaminetetraacetic acid and counted using a hemocytometer. Antibodies Melanocyte glycoprotein antigens transferrin receptor, human leukocyte antigen class I protein, melanotransferrin, melanoma proteoglycan, CALLA (gplOO), gpl80, gp45, gpl40/30, gpl30/100/27/24, gpllO, gpl 40 (integrins), gpl30, gp57, melanosomal membrane glycoprotein (CF21), lysosomal membrane protein (LAMP-1), and antibodies to these proteins have been described earlier [3,9,20-25]. Mouse monoclonal antibody (MoAb) TA99 has been described earlier [3]. Polyclonal antihodies to denatured human gp75 were made in rabbits at Cocalico Biologieals Inc. (Reanistown, PA); gp75 was purified from human melanoma cell line SK-MEL-19, as described earlier [10]. Two New Zealand White rabbits were injected with 25 /^g heat-denatured gp75 in complete Freund's adjuvant. Two weeks later, 15 ^ g gp75 was injected with incomplete Freund's adjuvant, followed by two more injections at intervals of 3 weeks. During the immunizations, the rabbits were test bled to monitor anti-gp75 response in the sera, and 2 weeks after the last injection, 30 ml of serum was collected and stored in aliquots at — 20°C. Serologic Assays For enzyme-linked immunosorbent assay (ELISA) to measure the expression of melanosomal antigens gp75 and CF21, 4 X lO"* cells were plated in Bat-bottom 96-well microtiter plates (Costar). After allowing 16-24 h for complete attachment, we added HMBA to a final concentration o f l , 3, or 5 mM, and medium was replaced every second day. Each day for tip to 5 d, cells in a set of plates were washed with phosphate-buffered saline (PBS), fixed with methanol:acetone (1:1) at -20°C for 10 min, washed several times with PBS, and blocked with 5% gamma-globulin—free bovine serum in PBS for 0.5-1.0 h. The plates were either used immediately for immunoassay or stored at 4°C until further use. In triplicate wells, 50 /xl of MoAb TA99 or CF21 serially diluted in gamma-globulin-free bovine serum in PBS was added and incubated at room temperature for 1-2 h. ELISA was performed as described earlier [10]. Wells that did not receive any antibody were used as reagent blanks, and several wells on each plate that received only alkaline-phosphatase—conjugated second antibody and the substrate were used to measure nonspecific antibody binding and endogenous alkaline phosphatase activity. Mixed Hemadsorption Assay Expression of cell surface antigens was assayed hy the mixed hemadsorption assay. Briefly, 200-500 cells in 10 |Li.l were plated in microtiter plates (Robins Scientific). After cell attachment for 16—24 h, 5 /nl of appropriate stock solution of HMBA in growth medium was added to obtain a final concentration o f l , 3, or 5 mM HMBA. Cells were grown at 37°C. Every second day, spent medium was aspirated using a syringe and needle and replaced with 10 /xl fresh medium with an appropriate final concentration of HMBA. On days 1, 3, and 5, the plates were washed once with PBS, and mixed hemadsorption assays with various antibodies were perfonned as described earlier [2]. Immunoprecipitation Metaholic (continuous and pulse/chase) labeling of cells in culture with ' ^S-methionine, cell lysis, and immunoprecipitation with MoAb TA99 were perfonned as described earlier [8]. Rabbit antimouse tyrosinase antibody (a generous gift from Dr. Bryan Fuller, University of Oklahoma Medical Scbool) followed by protein A Sepharose was used to precipitate metabolically labeled tyrosinase. Inimunoprecipitates were analyzed by sodium dodecylsulfate-polyacrylamide gel electrophoresis (Sf^S-PAGE) and visualized by fluorography, and immunoprecipitated proteins were quantitated by counting radioactivity in the regions of dried gels corresponding to the bands on the film.

THE JOURNAL OF INVESTIGATIVE DEKJVIATOLOGY

"Western Blot Analysis Cell pellets were washed thoroughly and lysed in lysis huffer containing 1% Nonidet P-40, 0.5% deoxycholate, and protease inliibitors leupeptin, pepstatin, aprotinin, and phenylmethylsulfonyl fluoride. Detergent-soluble protein was estimated using the BioRad D^. Protein Assay system (BioRad, Wchmond, CA). Twenty-five micrograms of protein from each lysate wa.s electrophoresed on 9% polyacrylamide gels and transferred to Immobilon (Millipore, Marlborough, MA), as described earlier [10]. The blots were then incubated for 1 h with rabbit anti-human gp75 serum diluted 1:1000, followed by incubation with peroxidaseconjugated goat anti-rabbit IgG diluted 1:20,000. Bound peroxidaseconjugated anti-rabbit IgG was visualized by chemiluminescence using the ECL detection kit (Amersham, Arlington Heights, IL). Northern Analysis Total cytoplasmie RNA was i.solated from cells in semiconfluent 1 50-cm~ flasks as described earlier [6]. Approximately 25 /xg of lUSIA was electrophoresed in 1% agarose gels containing formaldehyde, blotted onto GeneScreen membrane (NEN Dupont) by capillary transfer, prehybridized according to the manufacturer's instructions, and hybridized with '^^P-laheled full-length cDNA probes for human gp75 and tyrosinase at 42°C [6,7]. The blots were washed with 2 X sodium citrate/sodium chloride buffer at room temperature for 5 min, then with 2 X sodium citrate/sodium chloride containing 1% SDS at 60°C for 30 min, and finally with 0.1 X sodium citrate/sodium chloride at room temperature for 30 min. The blots were exposed to film for autoradiography with one intensifying screen. Tyrosinase Assays Tyrosine hydroxylase activity in cell lysates was measured as described earlier [10]. Briefly, enzyme activity in 1% Triton X-100 lysates equivalent to 10^' cells was assayed in triplicate in a final reaction volume of 200 /xl containing 0.2—0.3 jixCi "^H-tyrosine at 37°C for 1 ll. After adsorption ofthe reaction mixtures to activated charcoal, counts released as ' H j O were measured in a liquid scintillation counter. Background counts obtained as •^H2O released in buffer control tubes without cell extracts were equal to or less than 10% of counts ohtained in tuhes with cell e.xtracts. Background counts were subtracted from counts obtained for each reaction tube, and the mean ± SD of triplicate values is shown. '

RESULTS

HMBA Induces Dendritic Extension and Multidcndritic Morphology in Melanoma Cells Human melanoma cell lines SK-MEL-19 and SK-MEL-23 clone 22 (clone 22) express phenotypic traits characteristic of the mature stage of melanocytic differentiation [2]. SK-MEL-19 cells are spindle-shaped with short dendrites and bave dark brown to black pigmentation, and clone 22 cells display spindle-shaped bipolar morphology with black pigmentation. Both SK-MEL-19 and clone 22 cells cultured in the presence of HMBA acquired a polydendritic morphology. In tbe presence of 1 or 3 mM HMBA, there was no detectable change in the growth of tbe two melanoma cell lines. In 5 mM HMBA, both cell lines exhibited a measurable decrease in growth rates. At the end of four population doublings, the mean doubling time for SK-MEL-19 cells increased from 29 h in control conditions to 49 h in 5 mM HMBA. For clone 22, the mean doubling time increased from 23 h to 47 h in 5 mM HMBA (data not shown). No further decrease in the growth rate was apparent over long periods of incubation, even when 5 mM HMBA was present in the culture medium continuously for 3 weeks (mean population doubling time for both cell lines increased from 26 h to 57 h, i.e., a twofold increase). Do^vn-Regulation of gp75 (the b Locus Protein) in Melanoma Cells: A Specific Response to HMBA We studied the effect of HMBA on tbe expression of gp75 by ELISA and by immunofiuorescence staining using MoAb TA99. In both SKMEL-] 9 and clone 22 melanoma cells cultured in the presence of HMBA, tbere was a decrease in gp75 expression. Tbe effect was less! pronounced in SK-MEL-19 cells and sbowed no clear relation to the concentration of HMBA. In clone 22, the decrease in expression of gp75 was dependent on the concentration of HMBA. After 5 d in tbe presence of 5 mM HMBA, gp75 expression in clone 22 cells was almost completely extinguished (Figs 1, 2). In all seven melanoma cell lines tested, including tbe mouse melanoma B16, treatment with HMBA restilted in a 4O> ' (> to 90% decrease in gp75

VOL. lO.'i. N O . 1 JULY 1993

REGULATION OF gp7.S IN HUMAN MELANOCYTE CELLS

SK-MEL-19 tn

2.0 -1

clone 2 2

Table I.

DifFerential Regulation of gp75 and Tyrosinase

in Melanoma CeUs by HMBA'

0.3 -1

% Change

o

1.5 H 0.2 -

o C

o

1.0 0.1 -

0.5 -

O

10^ 10^ 10* 10^ 10^

115

10^ IO-' 10* 10^ 10^

1/TA99 antibody dilution Figure 1. Effect of HMBA on gp75 expression, SK-MEL-19 (/(•/?) and clone 22 (right) melanoma cells grown in 96-well plates were fixed and penneabilizod, and gp75 wa.v measured by ELISA u.fing MoAh TA99. Data are represented as antibody titration curves for untreated control cells (open liiangles) and cells treated with HMBA (5 mM for 5 d) (closed triangles). Each data point represents values obtained from four wells. Eiror bars represent ± SD.

synthesis (see below and Table I). W e chose SK-MEL-19 and clone 22 cell lines for detailed analyses. HMBA did not induce any detectable change in the expression of 16 other glycoproteins expressed by melanocytes and melanoma cell lines SK-MEL-19 and clone 22, These included cell surface proteins (transferrin receptor, integrins, melanoma proteoglycan, and a melanocyte-specific cell surface antigen) and intracellular membrane proteins (melanosomal CF21 and lysosomal membrane glycoprotein, LAMP-1) (data not shown). Tbere was no decrease in total cellular protein syntliesis or glycoprotein synthesis induced by HMBA (as measured, respectively, by incorporation of''^S-metbionine and " H-glucosamine by clone 22 cells into newly synthesized proteins). These data showed that HMBA specifically regulated expression of gp75, a protein tbat is intimately associated witb pigmentation, a differentiated pbenotype of melanocytic cells, HMBA-Induced Down-Regulation of gp75 Expression Occurs at the Single Cell Level and Is Reversible The decrease

in synthesis and expression of gp75 measured in bulk cultures could be a result of either down-regulation of gp75 syntbesis by HMBA at tbe individual cell level or selective expansion of a subpopulation of cells that express little or no gp75. To examine tbese possibilities, we studied gp75 expression in clone 22 cells by immunofluorescence micro.scopy using MoAb TA99. As shown in Fig 2, detectable but varying levels of gp75 expression could be seen in almost all cells of clone 22. In the presence of 3 mM HMBA for 5 d, only a very small population of cells showed weak to moderate staining

Cell Line

Synthesis

Tyrosinase Activity

SK-MEL-19 SK-MEL-21 SK-MEL-30 SK-MEL-188 B16 (mouse) SK-MEL-93 (2) SK-MEL-23 (clone 22)

-85 -88 -75 -56 -60 -39

+ 14 + 56 +33 +35 + 69 -58

-88

-40

" Tyrosinase activity and gp7.S syiithcsi.s were measured in various melanoma cell Hues grown in the presence of 5 uiM HMBA for .5 d. Tyrosiue hydrolyase activity was measured as release of "^H2O, aud gp7S synthesis wa.s estimated by metabolic labeling with "^'^S-metliioniue, as described iu tbe legeud for Figs 4 aud 5. Data are shown as percent enzj7iic' activity or synthesis coinp,irc'd vvith the untreated control of each cell line.

for gp75 expression. In 5 mM HMBA for 5 d, there was almost no detectable gp75 staining. This pattem of loss of gp75 expression, related to time and HMBA concentration, showed that HMBA down-regulated gp75 expression at tbe individtial cell level. In support of this notion, HMBA-induced down-regulation of gp75 was reversible, Witbin 24 li of removal of HMBA, a moderate to intense re-expression of gp75 could be seen in gp75-negative cells (Fig 3). This rapid reappearance of gp75 expression occurred uniformly in the entire population within a short period (12-24 b) with no detectable increase in cell number. These results sbowed that down-regulation of gp75 syntbesis by HMBA occurred at tbe individual cell level and not simply by selective expansion of a subpopulation of gp75-negative cells. Pigmentation and Tyrosinase Activity: Heterogeneity in

Response to HMBA Clone 22 and SK-MEL-19 cells showed a distinct response to HMBA with respect to melanin pigmentation. After 5-7 d of treatment witb HMBA, clone 22 cells became ligbtly pigmented, with occasional completely depigmented cells. On the other hand, HMBA induced an increase in pigmentation in SKMEL-19 cells compared with untreated cells. Examination of pellets of clone 22 and SK-MEL-19 cells treated with HMBA showed, respectively, lighter and darker pigmentation (data not showTi). Tyrosinase, which catalyzes tbe initial steps of tyrosine hydroxylation and oxidation of dibydroxyphenyialanine, is the critical and rate-limiting enzyme required for melanin syntbesis. To investigate wbether tbe observed differences in pigmentation were due to an eifect of HMBA on tyrosinase activity, we measured t)'rosine bydroxylase activity of tyrosinase in lysates of SK-MEL-19 and

Figure 2. HMBA down-regulates gp75 expression in differentiated melanoma cells. Clone 22 melanoma cells were grown on multi-well chamber glass slides witliout (yl) or with 1 mM (B), 3 mM (C,D), or 5 niM HMBA (E,F} for 5 d. Cells were fixed, permeabilized, and stained for gp75 expression and photographed. C,E, phase contrast micrographs of fields corre.sponding to areas in D and F. Bar, 250

116

VIJAYASAIUVDHI ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

B SK-MEL-19 0

1 3

Clone 22 5

0

1

3

Clone 22 5 mM HMBA

C

H

SK-MEL-19 C

H

Figure 5. De novo gp75 biosyntbesis is inbibited by HMBA. A, SK-MEL-19 and clone 22 cells grown in 1, 3, and 5 mM HMBA wore metabolically labeled vrith ^^^S-mothionine for 90 min on day 5. Newly synthosized gp75 was immunoprecipitatod using MoAb TA99 and analyzod by SDS-PAGE. B, tyrosina.se was immunoprocipitatod from SK-MEL-19 and clono 22 colls using a rabbit anti-mouso tyrosina.so antibody aftor 4 h of metabolic laboling. C, ly.sato from control colls grown without HMBA; H, lysates from cells grown in 5 niM HMBA for 5 d.

Figure 3. Down-regulation of gp75 by HMBA occurs in individual cells and is reversible. Clono 22 molanoma cells woro grown in duplicato wolls of glass chamber slldo.s widiout (A) or with 1 niM (B,C), 3 mM (D,E), or 5 mM (F,G) HMBA for 5 d. On day 5, one .slide was fixod, and medium in wolls on tlio othor slido was roplacod with HMBA-froo modium, incubated for an additional 24 h, and then fixed (C,E,C). Permeabilized cells wore stained for gp75 expression witb MoAb TA99. Bar, 250 ^m.

clone 22 cells treated with 1, 3, and 5 mM HMBA (Fig 4). In SK-MEL-19 cells, all three concentrations of HMBA caused a modest but reproducible increase (up to 30% compared with untreated cells) in tyrosinase activity. In contrast, in clone 22 cells, a 50% decrease in tyrosinase activity was ohserved as early as 24 h in 1 mM HMBA. No further decrease in the enzyme activity could he seen at higher concentrations of HMBA (up to 5 mM) or hy extending the time of exposure to HMBA up to 5 d (Fig 4). These efFects of HMBA on tyrosinase activity in melanoma cells were reproducihle in five independent experiments and were consistent with the observed eiFects of HMBA on pigmentation in individual experiments, i.e., a modest hut reproducihle increase in tyrosinase activity and darker melanin pigmentation of SK-MEL-19 cells and a decrease in enzyme activity and lighter pigmentation of clone 22 cells.

HMBA Inhibits Steady-State Accumulation of gp75 mRNA and Biosynthesis of gp75 We studied the biosynthesis of gp75 in untreated and HMBA-treated melanoma cells to understand whether modulation of gp75 hy HMBA was due to its effects on synthesis or stahility ofthe protein. After 5 d in 5 niM HMBA, very SK-MEL-19 clone 22 little or no new gp75 synthesis could he detected in either 150 -, SK-MEL-19 or clone 22 cells (Fig 5^). Inhihition of gp75 synthesis appeared to be regulated at the level of transcription or stnhiUty of gp75 mRNA. In the presence of 5 mM HMBA, the amount of gp75 100 < I mRNA was markedly decreased in hoth melanoma cell lines (Fig 6). We quantitated the inhihition of gp75 hiosynthesis hy measuring the radioactivity in gp75 protein hands. In clone 22 cells 50 cultured for 48 h in 1 mM HMBA, gp75 synthesis decreased to 30% of untreated control cells, whereas in SK-MEL-19 cells, comparahle inhihition was ohserved only aFter treatment with 3 mM HMBA for 5 d. Almost complete inhihition of gp75 synthesis could he seen in hoth cell lines cultured For 5 d in 5 mM HMBA. These results Days in c u l t u r e together with the data in Table I show that, although the degree oF inhibition varied, HMBA down-regulated de novo synthesis of gp75 in all seven melanoma cells lines tested. Figure 4. Tyrosinase activity in melanoma cells sbows variable response to HMBA. Tyrosino hydroxylaso activity was moasurod as Although synthesis of gp75 was efFectively inhibited hy HMBA in reloaso of ^H,O from 'H-tyrosino using lysatos of SK-MEL-19 (left) and hoth SK-MEL-19 and clone 22 cells, an amount of protein detectclone 22 (right) melanoma cells grown in the presence of 1 iiiM (circles), 3 able hy ELISA was present in SK-MEL-19 cells cultured in the mM (open triangles), or 5 mM (closed triangles) HMBA up to 5 d. The data presence of HMBA. This was in contrast to a complete depletion of shown aro reprosontativo of threo indo))ondent experiments. In control gp75 hy HMBA in clone 22 cells (Fig 1). This did not appear to he untroatod cells, tlio onzynic activity moasurod as counts in ^HjO ranged due to differences in the effect of HMBA on stability/half-life of from 200,000 to 400,000 cpni/mg protein/b. Enzymo activity in coll lysato.s newly synthesized gp75 hetween these cell lines. Pulse-chase at different times is shown as percent activity of control cells grown without HMBA. metabolic labeling experiments showed that HMBA did not cause

VOL. lO.'i. NO. 1 JULY

REGULATION OF Bp75 IN HUMAN MELANOCYTE CELLS

SK-MEL-19 0

1

3

Clone 22 5

0

1

3

5 mM HMBA

gp75

•HK

Tyr

— 28S

EtBr — 18S

117

steady-state tyrosinase mRNA levels in either SK-MEL-19 or clone 22 cells (Fig 6). These results suggested that HMBA regulated dc novo synthesis of tyrosinase in SK-MEL-19 and clone 22 cells with no marked effect on the accumulation of tyrosinase mRNA. Effect of HMBA on gp75 and Tyrosinase Expression: Implications for Differences in Pathways of Regulation We compared the effects of HMBA with those of other pharmacologic agents known to induce phenotypic changes in human melanoma cells to understand the possihle pathways of regulation of gp75 by HMBA. We measured gp75 expression and tyrosine hydrox-ylase activity in SK-MEL-19 aud clone 22 melanoma cells cultured in the presence of HMBA, 12-0-tetradecanoyl-phorhol-13-acetate (TPA), or cholera toxin alone; or in a comhination oFHMBA and TPA or HMBA and cholera toxin. In hoth SK-MEL-19 and clone 22 cell lines, cholera toxin added alone induced a two- to fourfold increase in gp75 expression. TPA had no effect on gp75 expression. When added together with cholera toxin or TPA, HMBA was still able to down-regulate gp75 expression in hoth cell lines (Fig 7A). SK-MEL-19 and clone 22 cells showed distinct responses in tyrosinase activity to HMBA, and to HMBA in comhination with cholera toxin. When added alone, HMBA induced a 1.5-2-Fold increase in SK-MEL-19 tyrosinase and caused a 50% decrease in clone 22 tyrosinase activity. Cholera toxin had no effect on HMBA-induced tyrosinase activity in SK-MEL-19 cells (Fig 7B). In clone 22, down-regulation oF tyrosinase hy HMBA was effectively hlocked when cholera toxin was added to the medium together with HMBA. These results showed that gp75 expression in hoth melanoma cell lines was regulated hy pathways that can be inhihited hy HMBA. The iiihihitory effect oFHMBA on gp75 was dominant over cAMP-induced up-regulation oF gp75 expression. These results also suggested that in mature melanoma cell lines, gp75 expression was regulated hy similar pathways, whereas tyrosinase expression was regulated by at least two distinct pathways.

Figure 6. HMBA specifically decreases tbe steady-state gp75 m R N A in melanoma cells. Twenty-five micrograms of total cytoplasmie RNA isolated from cells grown in medium alone (0) or in 1 niM, 3 niM, or 5 mM HMBA for 5 d was electrophoresed on denaturing agarose gels, transferred to GeneScreen, and probed, sequentially, with random primelabeled full-length cDNAs for tyrosinaso (Tyr) and gp7.S. It should be noted that the apparent increase in signal for tyrosinase in clone 22 cells treated with 3 niM HMUA is due to a slightly larger amount of RNA loaded in the lane marked i mM HMBA on tlic blot containing RNA from clone 22 cells (as sbown by etbidium bromide |EtBr| staining).

any measurahle change in the stahility/half-life of newly synthesized gp75 in either SK-MEL-19 or clone 22 cells. One possihle explanation for the persistence of measurahle gp75 in SK-MEL-19 cells could be accunitilation and longer intracellular retention of melanosomes in SK-MEL-19 cells compared with clone 22 cells. Electron microscopic examination of the high-speed particulate fraction of spent culture medium showed pigmented melanosomes in the culture medium ohtained from clone 22 cells, whereas no pigmented melanosomes could he seen in medium ohtained from SK-MEL-19 cells (data not shown). This suggested that persistence of gp75 expression in HMBA-treated SK-MEL-19 cells is due to inefficient release of gp75-coiitainiiig pigmented melanosomes into the medium hy these cells. HMBA Has No Effect on tbe Steady-State mRNA Levels of Tyrosinase There was very little or no difference in the autoradiographic intensity of protein hands immunoprecipitated hy antityrosinase antibody from control SK-MEL-19 cells or cells treated with 5 mM HMBA For 5 d. In clone 22 cells, treatment with 5 mM HMBA resulted in a decrease in inimunoprecipitahle tyrosinase (Fig 5B). However, this decrease in tyrosinase did not appear to he due to a decrease in tyrosinase mRNA. HMBA did not affect

DISCUSSION Comparative studies of the expression of cell surface and intracellular markers by human melanomas and their nonnal progenitor, epidermal melanocyte, have allowed identification oF three classes oF melanoma cells that correspond to different stages in melanocyte differentiation [2]. The abiHty oFmelanoma cells to differentiate is apparent both in vivo and in vitro. Eirst, the presence oF populations oF melanoma cells at different stages oF differentiation within a single metastatic lesion recapitulates the remarkahle phenotypic heterogeneity seen in vitro [2]. Second, upon induction with activators oF PKC and adenyl cyclase, melanoma cells in vitro undergo differentiation. Eor instance, nonpigmented melanoma cells that are induced to differentiate maniFest increased melanin synthesis and coordinated up-regulation oF a Family oF proteins, including melanosomal glycoproteins [3]. Thus, regulation oF proteins associated with melanin synthesis is an integral part oF melanocytic differentiation. Tyrosinase and gp75, which are associated with melanin synthesis, are two oF the best-characterized melanosomal proteins [4-8]. Our data show that HMBA specifically down-regulates the expression oFgp75 in melanoma cells and that in mature melanocytic cells, gp75 can he regulated independently of pigmentation, tyrosinase, or other specific differentiation traits. hi some melanonia cells, regulation of gp75 and tyro.sinase appears to be tightly coupled, and inducers that modulate the expression of gp75 also modulate the expression and activity oF tyrosinase (Table I). This closely linked control of the two differentiation gene products is presumed to reflect the coordinated regulation of tyrosinase and gp75 during nonnal melanocytic differentiation, ln a suhset of melanomas represented by SK-MEL19, the regulation of these proteins is uncoupled, i.e., downregulation of gp75 and up-regulation of tyrosinase hy the same pliarmacologic agent (Table I). The absence oF a coordinated

118

VIJAYASARADHI BT AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

T Mr X10-3

C

H

T

+ H

CT CT

+ H

SK-I\/IEL-19

22

B

2000 -

a (U

O

O

Figure 7. Down-regulation of gp75 by HMBA is dominant over cholera-toxin-induced up-regulation of gp75 expression. A, Western blot analysis of gp7.5 expression. One hundred micrograms of detergentsoluble cellular proteins from ceils grown for 5 d in medium alone (C), cbolera toxin (CT) alone, or HMBA (H) alone, and in a combination of cbolera toxin and HMBA (CT + H), was electrophoresed on 9% SDSPAGE, transferred to Immobilon, and probed witli rabbit anti-gp75 antibody, and protein bands were detected by cbemiluminescence. B, Tyrosine hydroxylase activity in SK-MEL-19 (top) and clone 22 cells (bottom) was measured in lysates obtained from cells grown for 5 d in medium alone (C), witb cholera toxin (CT), HMBA (H), or in a combination of cbolera toxin and HMBA (CT + H). Data from triplicate assay tubes corrected for background counts witb buffer control are sbown. Eiror bars represent ± SD. The experiment was repeated three times and in each experiment, triplicate sample mean values for enzyme activity in cells treated with HMBA, cholera toxin, or HMBA plus ebolera toxin were significantly different relative to control (Student t test; p S 0.05).

regulation of these two gene products may represent an alternative pathway within late-stage melanocyte differentiation. Altematively, these differences could he a result oF events occurring during tumor progression and, thereFore, unrelated to the melanocyte differentiation program. Given that tyrosinase and gp75 can he regulated independently hy HMBA in most melanoma cell lines tested, we Favor the Former possihility. The effects oFHMBA on a number oFmurine and htiman tumor cells in vitro have heen investigated. Induction oF terminal differentiation oF murine erythroleukemia cells is hy Far the most extensively studied effect oFHMBA oFcancer cells. In this model, HMBA has been shown to modulate speciFically the expression oF genes regulating cell proliFeration, as well as genes associated with the terminally differentiated phenotype oF erythroid cells (e.g., glohin genes) [26,27]. Our data show that HMBA specifically modulates the expression oF tyrosinase and gp75 in mature melanoma cells. We noted that down-regulation oFgp75 and tyrosinase hy HMBA was not accompanied hy the re-expression oF proteins that define early stages oF melanocyte differentiation. Although HMBA up-regulated tyrosinase in SK-MEL-19 cells, HMBA did not induce tyrosinase or alter the expression oF early melanocyte differentiation markers in nonpigmented, early melanomas (Vijayasaradhi ct al, unpuhlished observations). Thus, specific downregulation oF gp75 and differential regulation oF tyrosinase hy HMBA in mature melanomas shows that these traits can be regulated within the mature stage of melanocytic differentiation and do not represent a reversal to a less differentiated phenotype. PKC has heen shown to be involved in the pathway oF HMBAinduced differentiation in murine erythroleukemia cells [28,29], Agents that deplete PKC or prevent the conversion oF the membrane-hound Form to the soluble Form oF PKC block HMBAinduced differentiation oF murine erythroleukemia cells [26]. In SK-MEL-19 melanonia cells, the effects oFHMBA on tyrosinase were hlocked by TPA, consistent with a possible involvement oF the PKC-mediated pathway For tyrosinase regulation (data not shown). However, in clone 22 cells, cholera toxin (which causes elevation in intracellular cAMP levels), hut not TPA, hlocked the inhibitory effect oFHMBA on tyrosinase expression. These ohservations are consistent with the effects oF cAMP agonists, such as melanocyte-stimulating hormone, on tyrosinase activity hi mouse melanoma cells and support the notion that tyrosinase is regulated hy multiple pathways [4,30,31]. It is interesting that HMBA not only hlocked cholera toxin-induced stimulation oFgp75, hut was ahle to extinguish completely gp75 expression even in the presence oF cholera toxin. Thus, HMBA inhibition oF gp75 expression is dominant over cAMP-mediated pathways that up-regulate gp75 expression. These results highlight the differences in the regulation oF tyrosinase and gp75.

IVc thank Dr. I'anI Marks, Dr. Victoria lUclton, and Dr. Paul Chapman for the helpful snggestioiis and critical reading of the ntaiinscript. This work n>as sttpported hy National Institntes of Health grants AR41465 (SV) and CAS 6 82103 (ANH).

1.

Le Douariii NM: Cell lino segregation during peripheial nervmis system ontogeny. .SricMrc231:L';i.S-I.S22. 1986 2. Honghton AN. Eisinger M, Albino Al', CairnerossJG. Old LJ: Surfaee antigens of nic'Ianocytcs and melanomas. Markers of melanocyte differcntiatiim and melanoma subsets. / /;.v;i Med 1 56:17.S.';-1766. 1982 i. Hougliton AN, Real FX, Davis LJ. Cordon-Cardo C. Old LJ: Phenotypic hetorogiMieity of melanoma. Relation to the differentiation program of melanoma cells. J L.v/J Med 165:812-829. 1987 4. Halaban R, l'omerantz SH. Marshall S, Lambert DT. Lerner Ali: Regnlation of tyro.sinase in human melanocytes grown in culture. / Cell Biol 97-480-488, 198.1 .S. Kwon BS, Hat] AK. Ponierantz SH. Halaban R: Isolation and sequence of a clDNA clone for human tyrosinase tliat maps at the mouse c-alhino locus. Proc Natl Acad Sci USA 84:747.'i-.7477. 1987 6. Bouchard B. Fuller DB, Vijayasaradhi S, Houghton AN: Induction of pigmentarion in mouse fihrohla.sts by expression of human tyrosinase cDNA. J Exp Mal 169:2029-2042, 1989 7. Vijayasaradhi .S, Bouchard B. Houghtoii AN: The melanonia antigen gp7.S is the

VOL. 105. NO. 1 JULY 199.S

_; . 8. 9.

10. 11. 12. 13. 14. 15.

16. 17. 18.

19.

20.

hnmiui homologuc of the mouse b (brown) locus gene product. J Uxp Med 171:137.5-1380, 1990 Vijayasaradlii S. Doskoeb PM. Hougllton AN: Biosynthesis and intracellular movement of" the melanosomal membrane glyeoprotein gp7.S, tlio human b (brown) locn.s product. Hxp Cell Res 196:233-240, 1991 Thonrsou TM. Real FX. Murakami S, Cordon CC, Old LJ. Hongbton AN: DifFerentiation antigens of melanoeytes and melanonia: analysis of melanosome and cell surface markers of human pigmented cells with monoclonal antibodies._/ Iin>est Drnnalol 90:4.S9-466. 1988 Vijayasaradhi S. Hougbton AN: Purification of an autoantigenic 7.S-kDa human melanosomal glycoprotein. hit J C
REGULATION OF gp75 IN HUMAN MELANOCYTE CELLS

119

HF, Old LJ: Surface antigens of melanonia and melanocytes defined by mouse monoclonal antibodies: specificity analysis and comparison of andgen expres.sion in cultured cells aud ds.sues. Caucer Res 45:4401-4409, 1985 21. Retdg WJ. Murty VVVS. Mattes MJ. Cbagand RSK, Old LJ: Extracellular matrix-modulated expression of human cell surface glycoproteins A42 and J143. J E.\7) Med 164:1581-1599, 1986 22. Real FX. Carrato A, Schuessler MH. Welt S. Oettgen HF: IFN-gammaregtilated expression of differentiation antigen of btiman cells. J Immunol 140:1571-1576. 1988 23. Kantor RSR. Mattes MJ. Lloyd KO, Old LJ. Albino AP: Bioebemical analysis of two cell snrface glycoprotein compfexes. very common antigen 1 and very common antigen 2. / Biol Chem 262:15158-L5I65. 1987 24. Albino AP. Lloyd KO. Ikeda H. Old LJ: Biochemical analysis of a 130,000 molecnlar weight glycoprotein on human melanoma cells. J Immuuol 181: 159.5-1599. 1983 25. Mane SM, Marzella L. Bainton DF. Holt VK. Cha Y. Hildredi JEK. August JT: Purification and cliaracterization of buman lysosomal membrane glycoprotein. Arcb liiocliem Biopbys 268:360-378. 1989 26. Sbeircry M, Rifkind RA, Marks PA: Murine erythroleukemia cell differentiadon: DNase I hypersensitivity' and DNA mediylation near globin genes. Proc Ntitl Acad Sci USA 79:1180-1184. 1982 27. Ramsay RG. Ikeda K. Rifkind RA. Marks PA: Changes in gene expression associated with indnced differentiadon of erythroleukemia: protooncogenes. globin genes, and cell division. Proc Natl Acad Sci USA 83:6849-6853. 1986 28. Melloni E. Poiuremoli S. Micbetti M. Sacco O, Cakiroglu AG. Jackson JF. Rifkind RA, Marks PA: Protein kinase C activity and hexamethylene bis acetaniide-induced erj'tJiroJeukeniia cell differentiation. Proc Natl Acad Sci USA 84:5282-5286. 1987 29. Powell CT, Leng L, Dong L, Kiyokavva H. Bu.squets X, O'DriscoU K, Marks PA: Protein kinase C isozymes epsilon and alpba in murine erythroleukemia cells. Proc Natl Acad Sci USA 89:147-151, 1992 30. Hoganson GE, Ledwitz-Rigby F. Davidson RL. Fuller BB: Reguladon of tyrosinase uiRNA levels in monse uielanoma cell clones by melanocyte stimuladng hormone and cyclic AMP. Somat Cell Mol Genet 15:255-263. 1990 31.Jimenez M, Kameyania K, Maloy WL. Toniita Y. Hearing VJ: Manimahan tyrosinase: biosynthesis, processing, and modulation by melanocyte-stimnlating hormone. Proc Natl Acad Sci USA 85:3830-3834. 1988