[CANCER RESEARCH 51, 4815-4820, September 15, 1991]

Induction of Growth Factor RNA Expression in Human Malignant Melanoma:Markers of Transformation1

Anthony P. Albino,2 Brigid M. Davis, and David M. Nanus

Laboratory of Mammalian Cell Transformation [A. P. A., B. M. D., D. M, N.], Memorial Sloan Kettering Cancer Center, and Division of Solid Tumor Oncology¡D.M. N.], Department of Medicine, Memorial Sloan Kettering Cancer Center and Cornell University Medical College, New York, New York 10021

ABSTRACT

Alteration in the expression of growth factors is widely accepted asbeing one of several critical defects in the generation of the malignantcell. In the present study, 19 human metastatic melanoma cell lines werecompared to 14 normal human foreskin melanocyte cell lines for theproduction of RNA transcripts specific for 11 different growth factors.Using the extremely sensitive technique of polymerase chain reaction toamplify growth factor-specific complementary DNAs, we analyzed thefollowing: transforming growth factor (TGF) types a, fi,, ti¡.and /i.,,acidic(a) fibroblast growth factor (FGF), basic (b) FGF, FGF-5, keratinocytegrowth factor (KGF), HST, and platelet-derived growth factor (PDGF)types A and B. There were clear distinctions among the patterns ofgrowth factor RNA expression by normal melanocytes and malignantmelanoma cells. The prototypic melanocyte pattern of expression included TGF0!, TGF/Sj, and KGF. A subset of melanocyte cell lines alsoexpressed PDGFA transcripts. In contrast, melanoma cells characteristically expressed RNA transcripts of TGF0,, TGF&, TGF&, TGFa,bFGF, KGF, and PDGFA. Subsets of melanoma cell lines also expressedaFGF, FGF-5, and PDGFB. The results presented indicated that TGF/32,TGF«,and bFGF may be particularly important in melanomagenesis andthat these, as well as FGF-5, aFGF, and PDGFB, can be used as markersof transformation in this tumor type.

INTRODUCTION

Neoplastic transformation of human cells is consistentlyassociated with quantitative or qualitative alterations in theexpression of growth factors (1). These factors are thought toconfer a growth advantage upon the malignant cell by acting,in part, as autocrine and paracrine mediators of cellular proliferation. Dissecting the types of biologically important growthfactors produced by malignant and normal progenitor cells ofa particular tumor type is the first step in elucidating whetherparticular growth factors have any role in the diverse array ofphenotypic alterations observed in neoplastic cells. Malignantmelanoma provides a valuable model for study of stage-specificevents that accompany the progression of the melanocytethrough the transformation process (2). To date, numerousstudies have defined transformation-related alterations in (a)antigen expression (3,4), (b) cell surface growth factor receptorexpression (5), (c) production of growth factors (6, 7), (d)differentiation programs (8), (e) activation of oncogenes (9),and (/) normal chromosomal structure (10). Despite the factthat specific roles for each of these alterations in the development or maintenance of melanoma cells has yet to be elucidated,they have been useful as markers of progression, differentiation,and transformation of the melanocyte (11). In order to morecritically define genetic differences between the normal andmalignant melanocyte that may have etiological importance, in

Received 2/26/91; accepted 7/8/91.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This investigation was supported by USPHS Grant CA-37907 awarded bythe National Cancer Institute, Department of Health and Human Services.

2To whom requests for reprints should be addressed, at Laboratory of Mam

malian Cell Transformation, Memorial Sloan Kettering Cancer Center, 1275York Ave., New York, NY 10021.

the present study a series of cell lines established from humanmetastatic melanomas and normal human foreskin melanocyteswas analyzed for the production of RNA transcripts specific toa spectrum of growth factors. The results of this analysisindicated clear distinctions among the transcription patterns ofgrowth factors in normal melanocytes and metastatic melanoma cells.

MATERIALS AND METHODS

Cell Lines and Tumor Specimens. Cell lines established from humanmalignant melanomas (SK-MEL), human foreskin melanocytes, andhuman foreskin fibroblasts were derived and cultured as previouslydescribed (12). Purity of melanocyte cultures was determined by theexpression of melanocytic specific cell surface antigens, cytoplasmicmelanosomal proteins, and the presence of melanin (13). Melanomaand fibroblast cultures were maintained in Eagle's minimum essential

medium supplemented with 2 HIMglutamine, 1% nonessential aminoacids, 100 units/ml streptomycin, 100 units/ml penicillin, and 7.5%fetal bovine serum. Melanocytes were maintained in an identical medium supplemented with TPA3 (10 ng/ml) and cholera toxin (10~8 M)

(Sigma Chemical Co., St. Louis, MO) (14). RNA was extracted frommelanoma cells in culture from 8-40 passages and from melanocyteand foreskin fibroblast cell lines in culture from 2-10 passages. RNAextracted from a testicular germ cell tumor cell line, NTERA2 (15),was used as a positive control for HST transcription.

Preparation of RNA. Total RNA was extracted from logarithmicallygrowing cells by the thiocyanate/guanidinium method as describedpreviously (16). Extracted RNA was analyzed for integrity by agarosegel electrophoresis.

Preparation of cDNA. cDNA was generated from total RNA essentially as described in Ref. 17. Briefly, 1 pg of total RNA was incubatedat 68°Cfor 5 min, chilled on ice, and reverse transcribed in a final

volume of 20 ^1 containing: 50 mM Tris-HCl (pH 8.3), 75 mM KC1, 10mM dithiothreitol, 3 mM MgCl2, 50 MM each of dATP, dCTP,dGTP, and dTTP (Pharmacia, Piscataway, NJ), 500 ng random hex-amer primers (Pharmacia), 12 units RNA Guard (Pharmacia), and 200units MoMuLV reverse transcriptase (BRL Co., Gaithersburg, MD).The mixture was incubated at 37°Cfor 1 h, heated to 95°Cfor 10 min,and stored at -20°C.

PCR. Amplification was performed as described previously (9, 18).Briefly, cDNA was heated to 95°Cfor 10 min and cooled on ice for 5

min, and then 1 n\ was added to a 25-^1 reaction mixture containing2.5 ul lOx PCR reaction buffer [0.5 M KC1, 0.1 M Tris (pH 8.0), 15mM MgCl2, 0.1% gelatin (Fisher Scientific, Fair Lawn, NJ), 9.6 ^Meach dATP, dCTP, dGTP, and dTTP (Pharmacia Inc.), 75 ng of eachpriming oligomer, and 1.0 unit Taq polymerase (Perkin-Elmer Cetus,Norwalk, CT)] and H2O. Reaction mixtures were prepared for multiplesamples and aliquoted. A negative control consisting of a 25 /<!aliquotwithout the addition of cDNA was included in each amplification.Twenty-five //' mineral oil (Fisher Scientific) was layered over theaqueous phase to prevent evaporation. Amplification was performedusing a DNA thermal cycler (Perkin-Elmer Cetus) for 50 cycles. A cycleprofile consisted of 1 min at 95°Cfor denaturation, 1 min at 55-65°C

3The abbreviations used are: TPA, 12-O-tetradecanoylphorbol-13-acetate;

cDNA, complementary DNA; TGFa and TGF/3, transforming growth factor,types a and .(; aFGF and bFGF, fibroblast growth factor, types acidic and basic;KGF, keratinocyte growth factor; PDGFA and PDGFB, platelet-derived growthfactor, types A and B; UVR, UV radiation.

4815

on July 26, 2020. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GROWTH FACTOR PRODUCTION IN MELANOMA CELLS

for annealing, and 1.5 min at 73°Cfor primer extension. Optimal

annealing temperature was defined for each primer pair. Electropho-resis of 10 ¿ilreaction mixture on a 1.4% agarose (IBI, New Haven,CT) gel containing ethidium bromide was performed to evaluate amplification and size of fragments generated.

PCR Oligonucleotide Primers. Oligonucleotide primers were synthesized using a DNA synthesizer 380A (Applied Biosystems, Foster City,CA). Gene sequences used to construct Oligonucleotide primers werefrom published sources. All primer pairs used were designed to bracketcDNA sequences that in genomic DNA cross an intron-exon boundary.Table 1 details the specific Oligonucleotide regions used. Primer sequences were as follows: 0-actin, 5 '-GTGGGGCGCCCCAGGCACCA,3'-CTCCTTAATGTCACGCACGATTTC (19); TGF/3,, 5'-AAGTG-GATCCACGAGCCCAA, 3'-GCTGCACTTGCAGGAGCGCAC(20); TGF02, 5'-AAATGGATACACGAACCCAA, 3'-GCTGCATTT-GCAAGACTTTAC (21); TGFft, 5'-AAGTGGGTCCATGAACC-TAA, 3'-GCTACATTTACAAGACTTCAC (22); TGF«,5'-CGCCC-TGTTCGCTCTGGGTAT, 3 ' -AGGAGGTCCGCATGCTCACAG(23); aFGF, 5'-CTTCACAGCCCTGACCGAGAA, 3'-TGGCCAT-AGTGAGTCCGAGGA (24); bFGF, 5'-GTGTGTGCTAACCGTT-ACCT, 3'-GCTCTTAGCAGACATTGGAAG (25); FGF-5, 5'-AGT-GGCTTGGAGCAGAGCAG, 3'-TGGCTGCTCCGACTGCTTGA(26); KGF, 5'-ATCAGGACAGTGGCAGTTGGA, 3'-CATAGGAA-GAAAGTGGGCTGA (27); HST, 5'-ATCGGCTTCCACCTCCA-GGC, 3'-GTACTTGTAGGACTCGTAGGC (28); PDGFA, 5'-CC-CCTGCCCATTCGGAGGAAGAGA, 3 '-TTGGCCACCTTGACGC-TGCGGTG (23, 29); PDGFB, 5'-TGCTGAGTGACCACTCGATC,3'-TCCAAGGGTCTCCTTCAGTG (30).

Restriction Endonuclease Digestion of PCR Products. Amplified fragments were purified with Centricon-100 microconcentrators (Amicon,Danvers, MA), lyophilized in a Speedvac concentrator (Savant Instruments Inc., Hicksville, NY), and redissolved in water. Restrictionendonuclease digestion with the appropriate enzyme (5 units/^ig) wasperformed at 37°Cfor 4 h. Digested fragments were fractionated on a

3% NuSieve/1% SeaKem agarose (FMC Bioproducts, Rockland, ME)gel.

RESULTS

Analysis of PCR Products. Total RN A was extracted from 19cultured human metastatic melanoma cell lines and 14 normalhuman melanocyte cell lines. RNA was also extracted fromseven short-term cultures of normal human foreskin fibroblasts,which served as an independent control for these experiments.The RNAs were reverse transcribed into cDNA and then amplified using gene-specific primer pairs and polymerase chainreaction methodology. The location within each gene sequenceof all oligonucleotides used for PCR and the expected size ofeach amplified DNA fragment are shown in Table 1. Threedifferent methods were used to confirm the identity of the DNAproduct generated by sequence-specific primers: (a) PCR-am-plified DNA was analyzed by electrophoresis in agarose andcompared to fragment size predicted by the location of theprimers used (see Table 1 and Fig. 1); (¿>)PCR-amplified DNAwas first digested with restriction endonucleases and then fractionated by electrophoresis in agarose; the resultant fragmentswere compared to predicted sizes based on location of internalcleavage sites (method used for aFGF, FGF-5, HST, KGF, andPDGFB) (data not shown); (c) PCR-amplified DNA was se-quenced and the results compared to the known DNA sequenceof the growth factor gene (method used for /3-actin, TGF«,HST, and PDGFA) (data not shown). Fig. 2 illustrates restriction endonuclease analysis of the 247-base pair amplified PCRproducts generated from TGF0,, TGF&, and TGF/3, cDNA.Despite the high sequence homology of these genes, differencesin restriction endonuclease sites allowed the unambiguous de

tection of individual members of the TGF/3 family after PCRamplification with specific primers.

Growth Factor Production in Melanocytes and MelanomaCells. Table 2 summarizes the expression of RNA transcriptsof a panel of growth factors by 19 metastatic melanoma celllines and 14 melanocyte cell lines. All cell lines expressed RNAtranscripts for 0-actin, indicating the integrity of the cDNAsand their ability to serve as templates for amplification. Virtually all melanoma cell lines expressed RNA transcripts forTGF/3, (18 of 19), TGF(32 (18 of 19), TGFft, (19 of 19), TGFa(18 of 19), bFGF (19 of 19), KGF (18 of 19), and PDGFA (15of 19). In addition, a subset of cultured melanomas expressedaFGF (3 of 19), FGF-5 (5 of 19), and PDGFB (2 of 19). Nomelanoma culture (0 of 19) expressed HST-specific RNAtranscripts.

In contrast to melanoma cell lines, cultured melanocytes hada different pattern of expression for these growth factor RNAtranscripts. The typical melanocyte pattern appeared to beexpression of only TGF/3, (14 of 14), TGF/33 (14 of 14), andKGF (11 of 13). Cultured melanocytes had no detectable expression of TGFj32 (0 of 14), aFGF (0 of 13), bFGF (0 of 14), FGF-5 (0 of 13), PDGFB (0 of 13), or HST (0 of 13). A subset ofmelanocyte cell lines did, however, express PDGFA transcripts(5 of 14). In addition, because of the extreme sensitivity of thePCR reactions performed, TGFa: transcripts were detectable in3 of 13 cultured melanocyte cell lines. In dilution experiments,in which equal amounts of melanoma and melanocyte cDNAs

Table 1 Growth factor-specific Oligonucleotide primers for PCR

The nucleotide location within the sequence of the designated gene used forthe amplification of cDNA and the expected size of each amplified fragment.Sequences are from published sources (see "Materials and Methods'").

Location of Oligonucleotide in nucleotidesequence of cDNA

RNAtranscript0-ActinTGF0,TGF02TGF/3,TGFaaFGFbFGFFGF-5KGFHSTPDGFAPDGFB5'-Oligonucleotide103-122928-9471(1X1

11(1(11244-126358-7824-44229-248214-233298-318516-535622-64594-1133'-Oligonucleotide642-6191173-11531326-13061489-1469297-277425-405465-445672-653573-553764-744848-826719-700Fragment

size(basepairs)541247247247241403238460277250228627

Growth Factor Specific PCR Products

107X

603

310

194

Fig. 1. Predicted fragment sizes. PCR products of predicted size (see Table 1)were generated using sequence-specific primer pairs for each growth factor shown.Size markers (ordinate) are from OX174 DNA digested with //<«•!11endonuclease.

4816

on July 26, 2020. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GROWTH FACTOR PRODUCTION IN MELANOMA CELLS

|| TGFßl || TGFB2 || TGFB3 ||

1 2345678 9

61)3

310234194118

Fig. 2. Restriction endonuclease digestion. The 247-base pair PCR productswere generated from sequence-specific primer pairs for TGF/J|, TGFft, andTGF0J. Size markers (ordinate) are from 0X174 DNA digested with Hae\\\endonuclease. Uncut PCR products are shown in lanes l (TGFß,),4 (TGFft),and 7 (TGF/33). All three PCR products were digested with Rsal (lanes 2, 5, and«);Smal (lanes 3, 6, and 9). Rsal digests all three TGF0 PCR products butresults in different size restriction fragments (lanes 2, 5, and X). Smal digestsonly TCP/Si PCR product, resulting in 125- and 119-base pair doublet (lane 3).There are no Smal recognition sites in the TGFft and TGFfo PCR products asshown by the inability of this enzyme to reduce the size of the product (lanes 6and 9).

were titrated before PCR amplification, the results suggestedthat there was at least 10-fold more TGFa transcripts in melanoma cell lines than in TGFa-positive melanocyte cultures(data not shown).

Since melanocyte cell lines were grown in the presence ofTPA and cholera toxin, we examined melanoma cultures forinduction or suppression of growth factor-specific RNAs bymedium containing identical concentrations of TPA and cholera toxin as used for melanocyte growth. Representative mela

noma cultures, SK-MEL-93 DX-6 and SK-MEL-28, weregrown in medium supplemented with TPA and cholera toxinfor 2 weeks, at which time total RNA was extracted and usedfor PCR amplification with primers specific for a selected setof growth factors: TGFa, TGF/3,, TGF&, TGF&, bFGF, KGF,PDGFA, and PDGFB. The results showed no alterations in theexpression pattern of any of these growth factors when thesemelanomas were cultured in medium containing TPA andcholera toxin (data not shown). Thus, the action of TPA andcholera toxin on melanocytes does not include the suppressionof any of the growth factors examined.

Growth Factor Production in Foreskin Fibroblasts. Theexpression of RNA transcripts of these same growth factorswas examined in 7 human foreskin fibroblast cell lines. Thepattern of expression was different from that observed in eithermelanoma cell lines or melanocytes. With the exception ofPDGFB (0 of 7) and HST (0 of 7), all cultures expressed eachof the growth factors analyzed: TGF/3, (7 of 7), TGFfo (7 of 7),TGFft (7 of 7), TGFa (6 of 7), aFGF (7 of 7), bFGF (7 of 7),FGF-5 (7 of 7), KGF (6 of 7), and PDGFA (7 of 7). The failureto detect HST transcripts in melanoma cells, melanocytes, orfibroblasts was due to the lack of transcription of this gene inthese cells, and not to technical considerations, since HST-specific primers could detect HST transcripts in the germ celltumor cell line, NTERA2 (data not shown).

DISCUSSION

In the present study, we compared the production by culturedhuman malignant melanomas and normal melanocytes of RNAtranscripts for 11 different growth factors: TGFa, TGF/3,,TGF/32, TGF03, aFGF, bFGF, FGF-5, KGF, HST, PDGFA,and PDGFB. Clear distinctions in the patterns of growth factor

Table 2 Expression of growth factor-specific RNA transcriptsDetection of transcripts after PCR amplification of the designated gene in individual melanoma and melanocyte cell lines; +, expression; —,no expression.

Cell line Cell type 0-Actin TGF0, TGFft TGF& TGFa aFGF bFGFFGF-5SK-MEL-13

Melanoma + + + + + - +-SK-MEL-28Melanoma + + + + + - +-SK-MEL-30Melanoma + + + +-- ++SK-MEL-63Melanoma + + + + + - +-SK-MEL-81Melanoma + + + + + - +-SK-MEL-93

DX2 Melanoma + + + + + - +-SK-MEL-93DX3 Melanoma + + + + + - +-SK-MEL-93DX4 Melanoma + + -+ + + +-SK-MEL-93DX6 Melanoma + + + + + + +-SK-MEL-110

Melanoma + + + + + - ++SK-MEL-127Melanoma + + + + + + +-SK-MEL-136Melanoma + + + + + - +-SK-MEL-140Melanoma + + + + + - +-SK-MEL-146Melanoma + + + + + - +-SK-MEL-147Melanoma + + + + + - ++SK-MEL-186Melanoma + + + + + - ++SK-MEL-203Melanoma + + + + + - ++SK-MEL-204Melanoma + + + + + - +-MEWOMelanoma + - + + + - +-FSM4Melanocyte + +-+----FSM8WMelanocyte + +-+----FSM10Melanocyte + +-+----FSM11Melanocyte + +-+----FSM120Melanocyte + +-+----FSM138Melanocyte + +-+----FSM151Melanocyte + + - +----FSM152Melanocyte + +-+----FSM153Melanocyte + +-+----FSM196Melanocyte + +-+-"---FS215Melanocyte + + -+_»___FS220Melanocyte + +-+----FS225Melanocyte -1- + -+_"___FS238Melanocyte + + - + ND4 -ND"

Acknowledges detection of transcript after 50 amplificationcycles.*ND, not determined.HST

KGF PDGFAPDGFB-++—+ +——+ ++—+ +——+ ———+ ———— ++—+ ———+ ———+ +——+ +——+ +——+ +——+ +——+ +—-+ +——+ +——+ +——-1- +—-+ +-—+ ——__ __—— +——+ ———+ ———+ ———+ +——+ ———+ +——+ +——+ ———+ ——ND

ND - ND

4817

on July 26, 2020. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GROWTH FACTOR PRODUCTION IN MELANOMA CELLS

expression in these cell types were evident. Virtually all foreskinmelanocyte cultures expressed TGF/3i, TGFffj, and KGF RNAtranscripts, and 35% expressed PDGFA RNA transcripts. Astrikingly different pattern of expression was detectable inmelanoma cells. All melanoma lines expressed RNA transcriptsfor TGF/3,, TGF/32, TGFft, TGFa, bFGF, and KGF, whilemost also produced transcripts for PDGFA. A subset expressedaFGF-, FGF-5-, and PDGFB-specific RNAs. HST transcriptswere not detected in any melanocyte or melanoma cell line.Foreskin fibroblast cultures expressed TGFßj,TGF02, TGFßj,TGFa, aFGF, bFGF, KGF, FGF-5, and PDGFA but notPDGFB or HST.

Transforming Growth Factors. The TGF/3 polypeptide family,which consists of at least 3 distinct members (TGF/3i, TGF/32and TGFft) (20-22), can both inhibit and stimulate cell proliferation and can influence differentiation (31). TGF01 is ubiquitously expressed in human tissues, while expression of TGFj32and TGF/3, varies with cell type (20, 32). We showed thatcultured normal melanocytes and malignant melanomas express TGFßiand TGFfti but that only melanoma cells synthesize TGFj32 RNA transcripts. Since both normal and malignantmelanocytes express TGF/3, and TGF/33, it is possible that thesegenes are involved more in the normal growth and differentiation of melanocytes than in melanoma development. In contrast, the transcription of TGF/32 in all melanoma cell linesstudied indicates that this gene is regulated independently ofTGF/Ji and TGF/3j and may serve a critical role during thedevelopment of the melanoma cell. TGF/32 does have immu-nosuppressive effects (33), and it is speculated that inductionof immunosuppression in hosts by TGF/32-producing glioblas-tomas contributes to their growth by enabling these tumors toescape immune surveillance (34). TGF/3], which shares a widerange of biological effects with TGF/32 (35), has been shown toaffect the regulation of expression of class II histocompatibilityantigens in melanomas, and it is speculated that this regulationhas an adverse effect on host immune response (36). Thus, bycomparison with TGFßi,TGF/32 may similarly impact on thegrowth and maintenance of melanoma cells. While it remainsto be determined whether three TGF0 mRNAs are actuallytranslated into functional protein, the qualitative alteration inTGF02 transcription upon malignant transformation of themelanocyte may allow its use as a new marker of transformationin this tumor type. Whether the induction of TGF/32 expressionis associated with initial transforming events or with secondaryevents (e.g., those that affect growth, maintenance, or suppression of host immunity) is presently being analyzed.

TGF«is a 50-amino acid polypeptide that shares high structural homology in the receptor-binding domain with epidermalgrowth factor and can bind and activate the epidermal growthfactor receptor (37), thus stimulating proliferation by an auto-crine-mediated pathway. TGFa is expressed by a range ofmalignant human tissues (32, 38) and by some normal adulttissues (39). Our results showed that the majority of melanocytecell lines lacked expression of TGFa RNA. This confirms thefindings of a recent study that showed that TGF« protein isonly detectable in foreskin melanocytes after treatment withUVR (40). However, because of the extreme sensitivity of thePCR method, low level transcription of TGFa was detected ina few melanocyte cultures, suggesting that induction by UVRis quantitative as well. In contrast, virtually all cultures ofmetastatic melanomas produced large amounts of TGFa transcripts (at least 10-fold more transcripts than TGFa-positivemelanocytes). Thus, the expression of TGFa may be useful as

a marker of melanocyte transformation. While a specific rolefor TGFa in melanomagenesis remains obscure, TGFa is speculated to be an autocrine growth factor (32, 38), because itstimulates the growth of melanomas (40, 41) and melanocytes(40) in vitro. Moreover, the induction of TGFa by UVR maybe linked to the promotion of preneoplastic or neoplastic cells(40), and the presence of a restriction fragment length polymorphism in the TGFa gene of melanoma patients may be arisk factor (42).

Fibroblast Growth Factors. The FGF family is a diverse groupof polypeptide growth factors characterized by structural andsequence homology, heparin-binding characteristics, and theability to promote angiogenesis and provide mitogenic stimulifor a wide range of mesoderm- and ectoderm-derived cell types(43). The FGF family includes bFGF, aFGF, FGF-5, HST/K-FGF, FGF-6, Int-2, and KGF (43). bFGF, a cell-associatedpolypeptide, is found in a wide range of cell types and tissues(44). It can mediate neoplastic growth and the development ofmetastasis by acting as an autocrine mediator of cell proliferation or as a transforming oncogene (44). The induction of bFGFtranscription in all melanoma cell lines examined suggests thatit plays a fundamental role in the development of the malignantphenotype. This interpretation is strengthened by the observations that normal melanocytes require exogenous bFGF toproliferate in culture (45) and that the growth of melanomacells in culture can be inhibited by suppressing endogenousbFGF synthesis with antisense oligonucleotides (46). Althoughthe spontaneous production of bFGF by transformed melanocytes may allow unrestrained proliferation, it is unlikely to bean initiating event, since transfection of bFGF cDNA intomurine melanocytes abrogates proliferation dependence on exogenous bFGF but does not induce tumorigenicity (47).

KGF is a potent mitogen for keratinocytes but not fibroblasts(27, 48); unlike bFGF, it is efficiently secreted. KGF may benecessary for renewal of cell populations that line the skin (27).We show here that KGF transcripts were found in dermalfibroblasts, epidermal melanocytes, and malignant melanomacells. Whether the expression of KGF by melanocytes has anyrole in epithelial cell proliferation in the epidermis remainsunknown. However, since both normal and malignant melanocytes express KGF, it may be more related to the normal growthand differentiation of melanocytes and may not have any fundamental role in the development of melanoma.

Overexpression of normal FGF-5 and HST, which wereoriginally isolated from human neoplasias as oncogenes capableof transforming rodent cells (26, 49), can cause transformationof mammalian cells (28, 50). FGF-5 transcripts were not detected in any melanocyte culture, whereas approximately 25%of cultured melanomas did express FGF-5 transcripts. WhetherFGF-5 has a significant biological role in a subset of melanomasremains an open question, but the expression of FGF-5 only bymalignant melanoma cells indicates that the transcription ofthis gene may be a useful marker of transformation. Transcription of the HST gene, which may contribute to the developmentof breast cancer (51 ), appeared to be more tightly regulated incells of melanocytic lineage, since no melanocyte or melanomaculture expressed HST RNA transcripts. Thus, HST is unlikelyto play a role in the biology of melanoma.

Similar to bFGF, aFGF can act as an autocrine mediator ofcell proliferation and transform cells when overexpressed (52).A subset (15%) of melanoma cultures synthesized aFGF transcripts. While the biological impact of aFGF is as yet unknown,one conjecture is that aFGF is more relevant to late stages in

4818

on July 26, 2020. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GROWTH FACTOR PRODUCTION IN MELANOMA CELLS

tumor progression. This speculation is based on the fact thatthere were qualitative differences in aFGF expression amongfour cell lines established from separate metastatic deposits ofan individual patient. Two of these lines (SK-MEL-93 DX-4and SK-MEL-93 DX-6) expressed aFGF RNA, while two (SK-MEL-93 DX-2 and SK-MEL-93 DX-3) did not. The SK-MEL-93 DX-2 and SK-MEL-93 DX-3 cell lines were establishedfrom métastasesbiopsied 2 years before those used to establishSK-MEL-93 DX-4 and SK-MEL-93 DX-6, which appearedlate in the progression of this patient's disease (53).

Platelet-derived Growth Factors. PDGF, a potent mitogenand chemoattractant for fibroblasts (54), is found as homodi-mers of the A or B chains and as a heterodimeric AB structure(55). It has been shown that PDGF-related factors are producedby a subset of cultured melanomas but not by melanocytes (56),suggesting that PDGF production may be involved in melanoma development. However, because our results show thatapproximately 50% of melanocyte cell lines as well as 80% ofmelanoma cultures express PDGFA transcripts, PDGFAexpression is not an unequivocal marker of transformation inmelanoma. On the contrary, PDGFA may be more related tonormal growth and differentiation than to transformation. TheB chain of PDGF, which is structurally homologous to the c-sis protooncogene, has demonstrable transforming activitywhen permuted to the v-sis oncogene (57). Analyses of c-sis

expression at the RNA level have shown conflicting results.One study, using Northern analysis, could not detect PDGFBtranscripts in metastatic melanoma cells (56), but a more recentstudy, relying on RNase protection assays, showed that 60% ofmelanoma cell lines express PDGFB RNA transcripts (58).Using PCR methodology, we detected PDGFB transcripts inonly 10% of cultured metastatic melanomas (2 of 19). Since inour findings the frequency of expression is low, and since onlyone cell line of four established from four independent métastases from the same patient expresses PDGFB, it would appearthat the B chain of PDGF plays a minimal role in the etiologyof melanoma. However, the transcription of PDGFB in a subsetof melanomas could be useful as a marker of transformation.

In conclusion, the transcriptional activation of protoonco-

genes encoding growth factors and their receptors are hypothesized to have both direct and indirect roles in the regulationof neoplastic growth (59). The results of the present studysuggest that the development of the malignant melanoma cellis associated with the induction of TGF&, TGFa, and bFGF,while the induction of aFGF and FGF-5 may play a greaterrole in tumor progression than in tumorigenesis. The analysisof expression in precursor lesions and primary melanomas willbe of importance in determining the relevance of these growthfactors within the developmental pathway leading to metastaticmelanoma. The minimal amounts of RNA needed for analysisby PCR methodology should prove particularly useful in thestudy of these tissues, since it is usually difficult to obtainsufficient material for more conventional methodologies. Ultimately, the determination of consistent gene alterations inmelanomas may provide approaches for the better understanding of the biology of this neoplasia.

ACKNOWLEDGMENTS

We thank Harold McQuilla for expert technical assistance, Dr. EthanDmitrovsky for providing the NTERA2 cell line, and Dr. BerndSchmitz-Drager for useful discussions.

REFERENCES

1. Goustin, A. S., Leof, E. B., Shipley, G. D., and Moses, H. L. Growth factorsand cancer. Cancer Res., 46: 1015-1029, 1986.

2. Clark, W. H., Jr., Elder, D. E., Guerry, D., 4th, Epstein, M. N., Greene, M.H., and van Horn, M. A study of tumor progression: the precursor lesionsof superficial spreading and nodular melanoma. Hum. Pathol., 15: 11471165, 1984.

3. Houghton, A. N., Davis, L. J., Dracopoli, N. C., and Albino, A. P. Antigensexpressed by melanoma and melanocytes: studies of the immunology, biologyand genetics of melanoma. In: M. I. Green and T. Hamaoka (eds.), Development and Recognition of the Transformed Cell. pp. 373-383. New York:Plenum Publishers, 1987.

4. Elder, D. E., Rodeck, U., Thurin, J., et al. Antigenic profile of tumorprogression stages in human melanocytic nevi and melanomas. Cancer Res.,49:5091-5096, 1989.

5. Real, F. X., Rettig, W. J., Chesa, P. G., Melamed, M. R., Old, L. J., andMendelsohn, J. Expression of epidermal growth factor receptor in humancultured cells and tissues: relationship to cell lineage and stage of differentiation. Cancer Res., 46: 4726-4731, 1986.

6. Herlyn, M., Kath, R., Williams, N., Valyi Nagy, I., and Rodeck, U. Growth-regulatory factors for normal, premalignant, and malignant human cells invitro. Adv. Cancer Res., 54: 213-234, 1990.

7. Bordoni, R., Thomas, G., and Richmond, A. Growth factor modulation ofmelanoma growth stimulatory activity uiRN'A expression in human malig

nant melanoma cells correlates with cell growth. J. Cell Biochem., 39: 421-428, 1989.

8. Houghton, A. N., Real, F. X., Davis, L. J., Cordon Cardo, C., and Old, L.J. Phenotypic heterogeneity of melanoma. Relation to the differentiationprogram of melanoma cells. J. Exp. Med., 165: 812-829, 1987.

9. Albino, A. P., Nanus, D. M., Menile, I. R., et al. Analysis of ras oncogenesin malignant melanoma and precursor lesions: correlation of point mutationswith differentiation phenotype. Oncogene, 4: 1363-1374, 1989.

10. Balaban, G., Herlyn, M., Guerry, D. 4th, et al. Cytogenetics of humanmalignant melanoma and premalignant lesions. Cancer Genet. Cytogenet.,H: 429-439, 1984.

11. Albino, A. P., Houghton, A. N., Eisinger, M., et al. Class II histocompatibilityantigen expression in human melanocytes transformed by Harvey murinesarcoma virus (Ha-MSV) and Kirsten MSV retroviruses. J. Exp. Med., 164:1710-1722, 1986.

12. Albino, A. P., Le Strange, R., Oliff, A. I., Furth, M. E., and Old, L. J.Transforming rus genes from human melanoma: a manifestation of tumourheterogeneity? Nature (Lond.), 308:69-72, 1984.

13. Thomson, T. M., Real, F. X., Murakami, S., Cordon Cardo, C., Old, L. J.,and Houghton, A. N. Differentiation antigens of melanocytes and melanoma:analysis of melanosome and cell surface markers of human pigmented cellswith monoclonal antibodies. J. Invest. Dermatol., 90: 459-466, 1988.

14. Eisinger, M., and Marko, O. Selective proliferation of normal human melanocytes in vitro in the presence of phorbol ester and cholera toxin. Proc. Nati.Acad. Sci. USA, 79:2018-2022, 1982.

15. Andrews, P. W. Retinoic acid induces neuronal differentiation of a clonedhuman embryonal carcinoma cell line in vitro. Dev. Biol., 103: 285-293,1984.

16. Chomczynski, P., and Sacchi, N. Single-step method of RNA isolation byacid guanidinium thiocyanate-phenol-chloroform extraction. Anal.Biochem., 762: 156-159, 1987.

17. Sambrook, J., Fritsch, E. F., and Maniatis, T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor LaboratoryPress, 1989.

18. Saiki, R. K., Scharf, S., Faloona, F., et al. Enzymatic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sicklecell anemia. Science (Washington DC), 230: 1350-1354, 1985.

19. Ng, S. Y., Gunning, P., Eddy, R., et al. Evolution of the functional humanbeta-actin gene and its multi-pseudogene family: conservation of noncodingregions and chromosomal dispersion of pseudogenes. Mol. Cell. Biol., 5:2720-2732, 1985.

20. Derynck, R., Jarrett, J. A., Chen, E. Y., et al. Human transforming growthfactor-beta complementary DNA sequence and expression in normal andtransformed cells. Nature (Lond.), 316: 701-705, 1985.

21. Madisen, L., Webb, N. R., Rose, T. M., et al. Transforming growth factor-beta 2: cDNA cloning and sequence analysis. DNA, 7: 1-8, 1988.

22. Derynck, R., Lindquist, P. B., Lee, A., et al. A new type of transforminggrowth factor-beta, TGF-beta 3. EMBO J., 7: 3737-3743, 1988.

23. Rappolee, D. A., Mark, D., Banda, M. J., and Werb, Z. Wound macrophagesexpress TGF-alpha and other growth factors in vivo: analysis by mRNAphenotyping. Science (Washington DC), 241: 708-712, 1988.

24. Jaye, M., Howk, R., Burgess, W., et al. Human endothelial cell growth factor:cloning, nucleotide sequence, and chromosome localization. Science (Washington DC), 233: 541-545, 1986.

25. Abraham, J. A., Whang, J. L., Turnólo, A., et al. Human basic fibroblastgrowth factor: nucleotide sequence and genomic organization. EMBO J., 5:2523-2528, 1986.

26. Zhan, X., Bates, B., Hu, X. G., and Goldfarb, M. The human FGF-5oncogene encodes a novel protein related to fibroblast growth factors. Mol.Cell. Biol., 8: 3487-3495, 1988.

27. Finch, P. W., Rubin, J. S., Miki, T., Ron, D., and Aaronson, S. A. Human

4819

on July 26, 2020. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

GROWTH FACTOR PRODUCTION IN MELANOMA CELLS

KGF is FGF-related with properties of a paracrine effector of epithelial cellgrowth. Science (Washington DC), 245: 752-755, 1989.

28. Delli Bovi, P., Curatola, A. M., Kern, F. G., Greco, A., Ittmann, M., andBasilico, C. An oncogene isolated by transfection of Kaposi's sarcoma DNAencodes a growth factor that is a member of the FGF family. Cell. SO: 729-737, 1987.

29. Betsholtz, C., Johnsson, A.. Heldin, C. H., et al. cDNA sequence andchromosomal localization of human platelet-derived growth factor A-chainand its expression in tumour cell lines. Nature (Lond.), 320:695-699, 1986.

30. Rao, C. D., Igarashi, H., Pech, M. W., Robbins, K. C., and Aaronson, S. A.Oncogenic potential of the human platelet-derived growth factor transcrip-tional unit. Cold Spring Harbor Symp. Quant. Biol.. 51: 959-966, 1986.

31. Moses, H. L., Yang, E. Y'., and Pietenpol, J. A. TGF-ii stimulation andinhibition of cell proliferation: new mechanistic insights. Cell. 63: 245-247,1990.

32. Derynck. R., Goeddel, D. V.. Ullrich, A., et al. Synthesis of messenger RNAsfor transforming growth factors alpha and beta and the epidermal growthfactor receptor by human tumors. Cancer Res., 47: 707-712, 1987.

33. Siepi, C., Bodmer, S., Frei, K., et al. The glioblatoma derived T cell suppressorfactor/transforming growth factor-B2 inhibits T cell growth without affectingthe interaction of interleukin 2 with its receptor. Eur. J. luminimi.. 18: 593-602, 1988.

34. Bodmer, S., Strommer, K., Frei, K., et al. Immunosuppression and transforming growth factor-beta in glioblastoma. Preferential production of transforming growth factor-beta 2. J. Immunol., 143: 3222-3229, 1989.

35. Jennings. J. C.. Mohan. S.. Linkhart. T. A., Widstrom, R., and Baylink. D.J. Comparison of the biological actions of TGF beta-1 and TGF beta-2:differential activity in endothelial cells. J. Cell Physiol., 137:167-172, 1988.

36. Czarniecki, C. W., Chiù,H. H.. Wong. G. H.. McCabe, S. M., and Palladino.M. A. Transforming growth factor-beta 1 modulates the expression of classII histocompalibility antigens on human cells. J. Immunol., 140:4217-4223,1988.

37. Burgess, A. W. Epidermal growth factor and transforming growth factoralpha. Br. Med. Bull., 45: 401-424, 1989.

38. Imanishi. K., Vamaguchi. K., Suzuki, M.. Honda. S., Yanaihara. N.. andAbr. K. Production of transforming growth factor-alpha in human tumourcell lines. Br. J. Cancer, 59: 761-765. 1989.

39. Madtes, D. K., Raines, E. W., Sakariassen. K. S., Assonian, R. K., andSporn, M. B. Induction of transforming growth factor alpha in activatedhuman alveolar macrophages. Cell, 53: 285-293. 1988.

40. Ellem, K. A., Cullinan. M., Baumann, K. C.. and Dunstan, A. UVR inductionof TGF alpha: a possible autocrine mechanism for the epidermal melanocyticresponse and for promotion of epidermal carcinogenesis. Carcinogenesis(Lond.), 9/797-801, 1988.

41. Silici-lai \. S. E., Baker, F. L., Spitzer, G., et al. Biological effect of epidermalgrowth factor on the in vitro growth of human tumors. Cancer Res., 47:403-406. 1987.

42. Hayward, N.. Nancarrow, D.. Ellem. K.. Parsons. P.. and Kidson, C. A TaqiRFLP of the human TGF alpha gene is significantly associated with cutaneous malignant melanoma. Int. J. Cancer, 42: 558-561, 1988.

43. Goldfarb, M. The fibroblast growth factor family. Cell Growth Differ., I:439-445, 1990.

44. Rogelj. S., Weinberg, R. A., Fanning, P., and Klagsbrun, M. Basic fibroblastgrowth factor fused to a signal peptide transforms cells. Nature (Lond.), 331:173-175, 1988.

45. Halaban, R., Kwon, B. S., Ghosh, S., Delli Bovi, P., and Baird, A. bFGF asan autocrine growth factor for human melanomas. Oncogene Res., 3: 177-186, 1988.

46. Becker, D., Meier, C. B., and Herlyn, M. Proliferation of human malignantmelanomas is inhibited by antisense oligodeoxynucleotides targeted againstbasic fibroblast growth factor. EMBO J.. *: 3685-3691, 1989.

47. Dotto, G. P., Moellmann, G., Ghosh, S., Edwards, M., and Halaban, R.Transformation of murine melanocytes by basic fibroblast growth factorcDNA and oncogenes and selective suppression of the transformed phenotypein a reconstituted cutaneous environment. J. Cell Biol., 709: 3115-3128,1989.

48. Rubin, J. S., Osada, H., Finch, P. W., Taylor, W. G., Rudikoff, S., andAaronson, S. A. Purification and characterization of a newly identifiedgrowth factor specific for epithelial cells. Proc. Nati. Acad. Sci. USA, 86:802-806, 1989.

49. Bovi. P. D.. and Basilico. C. Isolation of a rearranged human transforminggene following transfection of Kaposi sarcoma DNA. Proc. Nati. Acad. Sci.USA, 84: 5660-5664, 1987.

50. Zhan, X., Culpepper, A., Reddy, M.. Loveless, J., and Goldfarb, M. Humanoncogenes detected by a defined medium culture assay. Oncogene, /: 369-376, 1987.

51. Adnane, J., Gaudray, P., Simon, M.-P., Simony-Lafontaine, J., Jeanteur, P.,and Theillet, C. Proto-oncogene amplification and human breast tumorphenotype. Oncogene, 4: 1389-1395. 1989.

52. Jaye. M., Lyall. R. M.. Mudd, R., Schlessinger, J., and Sarver. N. Theexpression of acidic fibroblast growth factor cDNA confers growth advantageand tumorigenesis to Swiss 3T3 cells. EMBO J., 7: 963-969, 1988.

53. Albino, A. P., Lloyd, K. O., Houghton, A. N., Oettgen, H. F., and Old, L. J.Heterogeneity in surface antigen and glycoprotein expression of cell linesderived from different melanoma métastasesof the same patient Implicationsfor the study of tumor antigens. J. Exp. Med., 154: 1764-1778, 1981.

54. Ross. R.. Raines, E. W.. and Bowen Pope, D. F. The biology of platelet-derived growth factor. Cell, 46: 155-169. 1986.

55. Shimokado. K., Raines. E. W.. Madtes, D. K., Barrett, T. B., Benditt, E. P.,and Ross, R. A significant part of macrophage-derived growth factor consistsof at least two forms of PDGF. Cell, 43: 277-286, 1985.

56. Westermark, B., Johnsson, A., Paulsson, Y.. et al. Human melanoma celllines of primary and metastatic origin express the genes encoding the chainsof platelet-derived growth factor (PDGF) and produce a PDGF-like growthfactor. Proc. Nati. Acad. Sci. USA, 83: 7197-7200, 1986.

57. Chiù, I. M., Reddy, E. P., Givol, D., Robbins, K. C., Tronick, S. R., andAaronson, S. A. Nucleotide sequence analysis identifies the human c-sisproto-oncogene as a structural gene for platelet-derived growth factor. Cell,37: 123-129, 1984.

58. Harsh, G. R., Keating, M. T.. Escobedo, J. A., and Williams, L. T. Plateletderived growth factor (PDGF) autocrine components in human tumor celllines. J. Neurooncol., 8: 1-12, 1990.

59. Slamon, D. J., deKernion, J. B., Verma, I. M.. and Cline, M. J. Expressionof cellular oncogenes in human malignancies. Science (Washington DC),224:256-262, 1984.

4820

on July 26, 2020. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1991;51:4815-4820. Cancer Res   Anthony P. Albino, Brigid M. Davis and David M. Nanus  Malignant Melanoma: Markers of TransformationInduction of Growth Factor RNA Expression in Human

  Updated version

  http://cancerres.aacrjournals.org/content/51/18/4815

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/51/18/4815To request permission to re-use all or part of this article, use this link

on July 26, 2020. © 1991 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from