Reduced Growth Factor Requirements and Accelerated Cell-Cycle Kinetics in Adult Human Melanocytes Transformed with SV40 Large T Antigen

Reduced Growth Factor Requirements and Accelerated Cell-Cycle Kinetics in Adult Hunian Melanocytes Transformed w^ith SV40 Large T Antigen Karoline Zepter, Andreas C. Haffiier, Uwe Trefzer, and Craig A. Elmets Department of Dermatology and Skin Diseases Research Center, Case Western Reserve University, Cleveland, Ohio, U.S.A.

Melanomas develop with high frequency in transgenic mice in which oncogenic sequences of the SV40 DNA tumor virus have been specifically targeted to melanocytes. To investigate the role of SV40 in melanomagenesis, cultured hunian melanocytes were transformed with a retroviral shuttle vector encoding the SV40 large T antigen and examined for changes in cell-cycle kinetics and grow^th-factor dependence. Colonies expressing the viral oncogene were morphologically indistinguishable from their non—T-antigen-transformed counterparts. Also like nomial melanocytes, the infected cells remained anchorage dependent and non-tumorigenic in nude mice. How^ever, T-antigen—positive cultures exhibited significantly accelerated population doubling times, increased saturation densities with highly confluent

monolayers and a three- to fourfold extended life span. Most interestingly, cell-cycle analysis revealed a measurable shift from quiescent to cycling cells in T-antigen—expressing cultures and an acquired ability to progress more rapidly through GI. Moreover, T-antigen-positive melanocytes proliferated in the absence of PMA and required markedly reduced levels of exogenous bFGF. These studies indicate that the viral oncogen of simian virus 40 provides melanocytes w^ith distinct grow^th advantages that may render these cells unusually susceptible to additional environmental challenges necessary for full expression of the malignant phenotype. Key words: oncogenel tumor suppressor geneslp53ltne1anotnalcell culture. J Invest Dermatol 104:755-762, 1995

E

of such key tumor-suppressor gene products as p53 and R b [7,11—14], the SV40 large T antigen may force the infected cell to increase nucleotide metabolism and DNA synthesis, thus providing an appropriate intracellular environment for its own replication [15]. The pleiotropic effects of large T-antigen expression on the growth, differentiation, and malignant potential of human epidermal melanocytes has received little attention [16,17]. This is somewhat surprising because melanocytes not only provide an excellent model system with which to examine positive and negative influences on cell growth and differentiation, but also have received considerable attention as a model for neoplastic development [17,18]. Moreover, in transgenic mice, susceptibility to cutaneous melanoma can be conferred by targeting the SV40 transgene to melanocytes with the tyrosinase promoter [19,20]. In the present study, we focused on proliferative features and cellcycle kinetics of SV40 large T-antigen—transformed adult human melanocytes, addressing the question of whether transfomiation of normal melanocytes with the SV40 large T antigen altered specific growth patterns and cell-cycle kinetics. Our findings indicate that the SV40 large T-antigen augments progression of melanocytes through the cell cycle and reduces their requirement for basic fibroblast growth factor (bFGF) an essential melanocyte mitogen. However, the cells do not undergo complete transformation to the malignant phenotype. These features suggest that introduction of SV4() large T andgen into melanocytes produces an intermediate stage between the nonnal and malignant cell, which may be useful for delineating key events in the melanomagenesis pathway.

pidermal melanocytes are subject to a highly complex network of heterogeneous activators and inhibitors that regulate their entrance and progression through the cell cycle [1,2]. Targets of these various stimuli are genetically determined positive (i.e., ras oncogene) [1,3,4] or negative (i.e., p53 and Rb tumor suppressor genes) [5,6] inducers of cellular proliferation and growth-factor expression. Mutations within these latter regulatory elements are particularly inclined to destabilize this complex system and may play a criticiil role in the pathogenesis of melanoma or its distinctive precursor lesions [1,3,5,6]. Similar states of imbalance between cell-cycle regulatory elements have been shown to occur when cells are transfected with certain viral genes with transfonning capabilities [4,7—10]. Because viruses depend on the host-cell machinery for their own development and multiplication, they may induce the expression of growth-regulating genes and may be particularly informative in revealing essential elements in the carcinogenesis process. The SV40 large T antigen, the dominant oncogenic protein expressed by the SV40 DNA tumor virus, is able to form stable, and functionally inactive, complexes with host-cell proteins that act specifically as proliferation inhibitors [7,11-14]. With inactivation

Manuscript rec(;ivcd September 2, 1994; revised December 1, 1994; ticccpted for publication January 20, 1995. Reprint requests to: Dr. Craig A. Elmets, Department of Dermatology, Case Western Reserve University, 11100 Euclid Avenue, Cleveland, OH 44106.

0022-202X/95/$09.50 • SSD10022-202X(95)00031-F • Copyright © 1995 by The Society for Investigative Demiatology, Inc. 755

756

ZEPTER ET AL

MATERIALS AND METHODS Cell Culture Adult melanocyte cultures from cadaver skin were established according to a protocol developed in our laboratory. Specimens of skin were incubated overnight at 4°C in 0.5% dispase (Boehringer Mannheim Biochemicals, hidianapolis, IN). Epidermal layers were peeled off and were further disaggregated in 0.15% trypsin (Sigma Chemical Co., St. Louis, MO) at 37°C. The resulting cell suspension, consisting mainly of keratinocytes and melanocytes, was passed through a tissue sieve and a fine needle to obtain single cells. Careful epidermal preparation and the utilization of serum-free medium prevented contamination with dermal fibroblasts. Primary cultures were established by placing the cell suspension in modified MCDB153 medium supplemented with 1 ng/ml bFGF, 0.4% (v/v) bovine pituitary extract (BPE), 10 ng/ml phorbol myristate acetate (PMA), 100 U/ml penicillin, 100 jug/ml streptomycin (all components firom Gibco, Grand Island, New York), and 2 mM CaClj. The relatively liigh calcium concentration led to terminal keratinocyte differentiation thereby selecting for melanocytes. Pure populations were usually obtained by the second passage. Melanocytes were fed biweekly and passaged every 6 - 8 d. The melanoma cell line SK-MEL 28 (American Type Culture Collection, Ardmore, OK) was maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum (FBS) (Hyclone Laboratories Inc., Logan, UT), 100 U/ml penicillin, and 100 p.g/nil streptomycin (Gibco BRL, Gaithersburg, MD). All microscopic photographs were taken witb a Nikon FX-35 DX camera attached to a Nikon UFX-DX microscope. Retroviral Transformation and Identification of Transformants The replication-defective retroviral vector pZIPTEX, encoding both the SV40 large T gene and the bacterial neomycin resistance gene, was based on the pZIP-Neo SV(X)L vector [21]. Tbe virus was transfected into '^I'CRIP cells carrying a mutant Moloney leukemia virus stably integrated into tbeir genome. Lacking tbe essential ^I' site to be packaged into virions, the helper virus itself cannot be propagated, but provides all obligate trans functions to produce bigh titers of infectious replication-defective retroviral particles. The transfected packaging cell line was kindly provided by Dr. J. Jacobberger [21]. Packaging cells were cloned in 800 fj-g/ml G418 (Gibco) containing medium prior to tbe transformation experiments to ensure retroviral integration. Retrovirus containing supernatant was collected 24 b after medium cbange, filtered tbrough a 0.22-/xm filter (Gelman Sciences, Ann Arbor, MI) and supplemented with 4 /xg polybrene/ml (Sigma) to facilitate virus attachment to tbe bost cells. 5 X 10^ melanocytes from the second passage were seeded on 10-cm petri dishes 36 b prior to transforination. Medium was replaced witb infectious supernatant and cells repeatedly incubated for 2 h at 37°C, adding fresh virus-containing medium each time. Forty-eight hours post-infection, melanocytes were exposed to 400 /Lig/ml G418 in culture medium and resistant colonies were expanded during tbe following weeks. Cell Fixation, Staining, and Flow Cytometry Flow-cytometric analysis was used to detect SV40 large T-antigen expression in transformed melanocytes and to determine the cell-cycle distribution in normal and T-antigen-positive melanocytes. One million cells per sample were fi.xed witb metlianol at — 20°C, incubated witb 1 jLLg of mouse anti-SV40 large T antigen antibody diluted in 50 /xl (PAb41 6, Oncogene Science, Manbasset, NY), followed by staining witb goat anti-mouse IgG F(ab')2 fragments conjugated to fluorescein (Boehringer Mannbeim). For DNA measurements, cells were placed in 150 [iX phosphate-buffered saline (PBS) containing RNase A (0.04 Kunitz units/sample) for 30 min. At the end of tbe incubation period 500 /xl of PBS containing 100 (xg/ml propidium iodide was added |21]. Analyses were performed on a flow cytometer (Cytofluorograpb fis; Ortbo Instruments, Westwood, MA), using tbe 488-nm line of an argon laser operating at 250 mW. Green fluorescence was collected tbrough a 53()/20-nni bandpass filter set, and red fluorescence was collected above 640 nm. A doublet discriminator (peak versus integrated signal) as the primary gate excluded cell aggregates. Immunofluorescence was logarithmically amplified and at least 10,000 cells were analyzed per sample. Tbe percentage of tbe total cell population present in GI or G2 + M was calculated from DNA histograms using tbe Cicero software (Verity Software House, Topsbam, ME). Results represent averages of at least tbree independently conducted experiments using passage-matcbed cells of dif-' ferent donors. Cell Cycle Inhibitors Apliidicolin (Sigma), a potent inhibitor of DNA syntbesis (S pbase), was dissolved in dimethyl sulfoxide (Sigma) at a concentration of 1.5 mg/ml and used at a concentration of 1 (xg/nil. Colcemid (Gibco), wbich jirevents completion of mitosis (G2/M pbase), was added to tbe medium at 0.2 /xg/ml. In preliminary experiments, tbe drugs were titrated to determine tbe minimal concentration necessary to yield tbe maximum fraction of cells in tbe indicated compartment.

THE JOURNAL OF INVE.STIGATIVE DERMATOLOGY

Grow^th Kinetics Proliferation of normal and T-antigen-positive melanocytes and melanoma cells under varying culture conditions (regarding cell deiisitry and bFGF supply) was assessed. For eacb condition, defined numbers of cells were seeded in triplicates in 24-well plates, were collected by trypsinization and were counted in a bemocytometer. Population doubling times were calculated from tbe slope of tbe log-phase curves. To quantify saturation densities in tbe respective cultures, cells were grown until n o further increase in confluence could be observed. Tbey were tben barvested by trypsinization and counted. Saturation densities are expressed as total numbers of cells/cm^ growtb surface. Melanin Content Relative melanin contents of tbe different cell lines were determined by beating pellets of lO'' cells per sample, in 1 ml 0.1 M N a O H , at 95°C for 90 min. Tbe optical density of the supernatants was measured at 475 nm witb a Beckinann DU 640 spectropbotometer and fibroblasts at similar cell numbers served as a blank standard. Northern Blot Hybridization Expression of tyrosinase mRNA was determined by a standard RNA bybridization procedure [22]. Ten micrograms of total cellular RNA from normal and T-antigen-transformed melanocytes and from melanoma cells was separated on a denaturing 1.4% agarose gel, transferred to a nylon membrane, and ultraviolet crosslinked. Tbe gene-specific probe consisted of a 1.6-kb cDNA clone of tbe buman tyrosinase gene inserted at tbe Eco RI site of a plasmid vector (pMEL34, a kind gift of R. Halaban [22]) and was labeled witb '^P using tbe random oligo priming metbod (Gibco BRL). D e t e c t i o n o f bFGF Protein Following detergent lysis (1% Triton X-lOO) of eacb 5 X 10'' cells, 30 /ng soluble protein sample was subjected to 1 Z% sodium dodecylsulfate—polyacrylamide gel electropboresis (SDSP A G P ) and electrotransfer on a nitrocellulose membrane. After blocking of nonspecific binding sites, tbe membranes were incubated witb a monoclonal rabbit anti-buman bFGF antibody (Sigma) and bands were visualized using tbe P A K O AEC-substrate system (DAKO, Carpinteria, CA). Anchorage-Independent Growth Six-well plates, covered witb 1 ml/ well 0.5% agar (Difco Laboratories, Detroit, MI) in culture medium, were overlaid witb 1 ml/well of a 0.3% agar suspension containing eitber 10'' or 10"* eells/nil and 10"A FBS. One milliliter of culture medium was added weekJy and tbe development of colonies was observed over a period of 3 weelts. Groups of more tban .30 cells were considered as colonies. Aggregates of cells present during tbe initial plating process were marked and disregarded. Tuniorigenicity in Nude Mice To evaluate malignant transformation, ]()'! f-antigen—expressing melanocytes in 100 /xl sterile Flank's balanced salt solution were injected subcutaneously into tbe subscapular region of 4_6.,week—old female atbymic mice (Animal Researcb Facility, C W R U ) . Control animals received Identical numbers of eitber normal melanocytes or tumor cells in tbe same preparation. Eacb group consisted of five mice. Animals were examined for tumor formation at regular intervals for 20 weeks.

RESULTS Isolation and Characterization of T-Antigen-Expressing Transformants SV40 large T-antigen-transformed melanocytes were isolated by incubating second passage cultures in the logarithmic growth phase with infectious, virtis containing supernatant. Within 10-14 d, drug-resistant colonies were clearly visible, further expanded ctt masse, and T-antigen expression monitored by fluorescence-activated cell sorter analysis. Only cultures that were 100'% positive for T-antigen expression were suhjected to experimental conditions. SlOO and DOPA stains confirmed the melanocytic origin of the transformants and the absence of keratinocytes and fibroblasts (data not shown). Although slightly larger with increased cell body volume, Tantigen-expres.sing melanocytes remained morphologically unchafiged upon transformation. Like their non-transformed counterparts, they displayed the typical features of cultured melanocytes witH bi- and multipolar dendrites (Fig 1). Melanoma cells, sucb as the spindle-shaped SK-MEL 28 line, frequently lose this characteristic appearance. Cultured human melanocytes displayed a characteristic growth pattern of regularly distributed cells that formed a network of dendritic contacts, but rarely overlapped and usually stopped dividing when cells reached 70 — 80% confluence. In striking contrast to this apparent contact inhibition, SV40 large Tantigen-transformed melanocytes formed tightly packed monolay-

VOL. 104, NO. S MAY 1995

SV40 LARGE T-ANTIGEN-TRANSFORMED MELANOCYTES

757

Figure 1. SV40 T-antigen-transformed melanocytes remain mor-

Figure 2. SV40 T-antigen expression increases the saturation den-

phologically unchanged. The pbotomicrograpbs sbow semicoiiflnent cultures of (a) normal human melanocytes at passage tbree and (ft) SV40 T-antigen—transformed melanocytes, second-passage post-transfection. Bars, 15 /xm.

sity o f human melanocytes. Cells were plated at 4 X 10' cells/cm* and continuously grown witbout passaging. The pbase pbotoiTiicrograpbs depict tbe typical appearance of ci) non-transformed melanocytes at saturation density witb an even distribution of cells that rarely overlapped, h) In contrast, T-antigen—positive melanocytes fonned tightly packed monolayers. Bars, 100 /nm.

ers with continuing proliferative activity. The cells within these hyperconfluent cultures were mostly bipolar and firequently overlapped (Fig 2). As confirmed by positive DOPA reaction, SV40 large T-antigentransfonned melanocytes continued to produce melanin, but, as determined by measurements of the relative optical densities of these pellets, showed 6.5 times lower absorption values compared to naive cells (optical density values: fibroblasts 0.008, SK-MEL28 0.009, T-antigen-transformed melanocytes 0.065, normal melanocytes 0.559). These findings corresponded to the results of Northem hlot analyses, which showed high levels of tyrosinase mRNA in non-transformed melanocytes and a significantly weaker signal in T-antigen—positive cells. No detectable tyrosinase mRNA was found in extracts of SK-Mel 28 melanoma cultures (Fig 3). S'V40 Large T Antigen Increases the Mitotic Fraction of Transfected Cnltnres and Accelerates Their Progress Through the Cell Cycle As calculated from third-passage cultures, the average cell yield after 8 d of culture was 15.9 ± 1.59 X 10'' cells/cm^ for T-antigen-positive cells compared to 5.8 ± 0.53 X 10* cells/cm for non-transformed melanocytes, indicating that the growth of SV40 transformed melanocytes was substantially enhanced. To determine whether SV40 large T-antigen expression influenced the percentage of mitotic cells within the population, the DNA distribution pattern in normal and transformed cultures at

varying densities was assessed. This was accomplished by plating cells and harvesting them after two population doubling times (i.e., 72 h for T-antigen—transformed cells and 96 h for nonnal melanocytes, respectively). In SV40 large T-antigen-positive populations, the mitotic fi-action (cells in G2/M) was elevated compared to normal melanocytes at a density of 2 X 10"* cells/cm"^ and remained essentially stable even at higher cell densities (Tahle I). In cultures of non-transformed melanocytes, the increase in population density was associated with a reduction of cells in the G2/M phase (Tahle I). Similar restilts were observed when cells were harvested after one and tliree doubling times (data not showti). This was confirmed by examining the cell-cycle distribution in non-synchronized cultures of the two different melanocyte populations (Tahle II). In those experiments, 68% of the normal population was found in GO/Gl (i.e., the quiescent phase), another 20'Vi) was in the S phase and the remaining \2'y« was undergoing mitosis (G2/M phase). In ctiltures of T-antigen—transformed melanocytes, however, almost twice as many cells (23'/o) displayed a mitotic DNA pattem at any given time (G2/M), 22% were progressing through the S phase, whereas the percentage of cells in the GO/Gl phase decreased to 55%. These experiments indicated that the number of SV40 large T-antigen—positive cells in mitosis was increased compared to normal melanocytes and remained at an elevated level even in more confluent cultures, implying that the

758

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

ZEPTER ET AL

II. SV40 Transformed Melanocytes from SemiConfluent, Non-Synchronized Cultures Contain More Cells in the GO/Gl Phase of the Cell Cycle than Do Normal Melanocytes" GO/Gl-Pbase melanocytes T-arltigen— expressing melanocytes

1.9 kb

a b Figure 3. Northern analysis of normal, SV40 T-antigen-transformed melanocytes and the SK-M£L 28 melanoma line shows a substantial reduction of tyrosinase mRNA levels in T-antigen— positive cells. mRNA of logarithmically growing cells at passage 3 (bigber passage number for tbe melanoma sample) was prepared using tbe guanidinium-isotbiocyanate/cesiuin chloride gradient metbod. Five micrograms of eacb sample was eltctropboresed on an 1.4% agarose gel and bybridized witb tbe radiolabeled DNA insert of pMEL 34 [23]. The autoradiograpb (72 h exposure) shows (a) normal human melanocytes, (fc) T-antigen— transformed melanocytes, and (c) SK-MEL 28 nielanoma cells.

normal constraints tbat restrict cell growth at high population densities was reduced. The time needed for T-antigen—expressing melanocytes to progress through the cell cycle was also accelerated, averaging 36 h compared to 49 h for normal melanocytes. In the transformants, the GO/Gl phase lasted approximately 18 ± 4 h (exit of 50% of cells from GO/Gl within 9 h) versus 30 ± 6 h in nonnal melanocytes (exit of 50% of cells from GO/Gl within 15 h). No differences were seen in the length of G2/M (9 ± 2 h and 10 ± 2 h for transformed and normal melanocytes, respectively) (Tahle III). From these studies, it was concluded that the reduction in the length of the cell cycle was caused primarily by a shortened GI phase in the transformants (Fig 4a, b). Transfortned Melanocytes Require Suhstantially Less hFGF and Proliferate Without PMA bFGF is an essential component Tahle I. SV40 Transformed Melanocyte Cultures Contain More Mitotic Cells and Display Less Contact Inhihition than Do Normal Melanocytes Cells in tbe G2/M Pbase of tbe Cell Cycle Number of Cells Seeded"

Normal Melanocytes

T-Antigen—Expressing Melanocytes

2 X 10' cells/cm^ 4 X 10' cells/cm^ 8 X 10' cells/cm^

17.4% 12.3% 1.9%

27.2%, 23.4%, 20.8%

G2/M-Phase

68%

for tnelanocyte mitogenesis [23,24]. Nonnal melanocytes lack the synthetic capacity for bFGF. Ratber they depend on its production by exogenous sources (keratinocytes, fibroblasts). In contrast, mela^noma cells characteristically both synthesize bFGF and respor>d to i' if 31^ autocrine manner. Normal melanocytes required 1 n g / m l purified bFGF to achieve a 50% increase in cell numbers over 10 d and lower concentrations (0 — 0.4 ng/ml) led to extensive cell death. T-antigen—transformation reduced the bFGF requirem e n t to 4()')^i (0.4 ng/ml) of the original level, suggesting endogenoiJS production of this indispensable growth factor. Tliis was confirmed by Western blot analysis. bFGF protein appeared as a band of 17 kd in samples of T-antigen-transfonned melanocytes and melanoma cells that was not detectable in normal melanocytes (Fig 5). Tbe band was more prominent in the melanoma cells than in T-antigen-transfonned melanocytes. In contrast to SK-MEL 28 cells, complete independence from exogenous bFGF could not be achieved in the T-antigen—transformed population (Figs 6 and 7). Normal melanocytes also require PMA for growth. Omission of PMyV from the growth medium results in their rapid senescence witb flattening and dendrite reduction within 48 h and population deatli vvithin 5 d. Melanoma cell lines, on the other hand, do not require this agent for proliferation. Studies were next conducted to detef"^''^^ whether T-antigen—positive melanocytes behaved more like riormal melanocytes or like melanoma cell lines in tliis regard. Large T-antigen-positive melanocytes displayed no reduction in theii' proliferative activity in the absence of PMA, although reduced deiicJrite formation was observed (data not shown). T-AJMtigen-Expressing Melanocytes Display an Expanded Dov»Wing Time, hut Remain Anchorage-Dependent and Not»-Tumorigenic in Nude Mice In normal human epidermis, mel^nocytes only rarely undergo mitosis and their life span in vitro is usiJally limited to 10 to 60 doublings depending on the age of the skin source [18,25]. T-antigen-transformed cultures of adult skinderived melanocytes, however, exhibited a substantial extension of their lif*^ span, at least quadrupling the number of possible passages. Lg^rge T-antigen-transformed melanocytes were evaluated for their capacity to undergo anchorage-independent growth as assessed by their ability to grow in soft agar. Like non-transformed mela.nocytes, they failed to form colonies in soft agar, whereas

III. Transformed Melanocytes Progress More Jtapidly Through the GO/Gl Phase of the Cell Cycle"

melanocytes

" In contrast to normal melanocytes, transformed cells remain highly proliferative even at increasing cell densities. Normal and T-antigen-expres,sin|; melanocytes were plated at the concentrations indicated, ;ind nfter 72 h, flow cytometry for DNA content was performed as detailed in Materials and Methods. Percentage values represent results from one experiment of three performed.

S-Pbase''

G O / G l Pbase

S-Pbase''

G 2 / M Phase

30 ± 6 b

9 ± 2b

10 ± 2 b

16 ± 4 h

1b

9 ± 2h

" j^jstribntion of cells in the difterent stages of the cell cycle: FACS analysis was i;Qjj(j^jCted in semiconfliient, logarithmically growing cultures of the same passage. Sampjes of 0.5 X 1 O*^* cells were prepared as described in Materials and Methods. Results repres*^"t averages of three independent experiments using cells of different donors. Data j-epresents the mean ± SD. '• fi^^ calculated from population doubling times in logarithmically growing cultures.

VOL. 104, NO. 5

MAY 1995

SV40 LARGE T-ANTIGEN-Tl^U^NSFORMED MELANOCYTES

759

% of population inG1

4

8

12

16

hours .^transformed melanocytes

.^normal melanocytes

% of population in G2/M

25 20 15 10 5

17 kd hours Hi-normal melanocytes

-^transformed melanocytes

Figure 4. In adult melanocytes, SV40 T-antigen transfomiation accelerates the passage through GO/Gl, hut does not influence the G2/M and S phase of the cell cycle. Tbe flow-cytometric DNA analysis quantifies cells ill the respective cycle pbase (percent of tbe total population) in samples taken at varions time points, a) Exit times (in liours) from t h e GI pbase of tbe cell cycle in cnltnres of normal (•) and SV40-transromied melanocytes (A) treated with Colcemid (cells blocked in G2/M). I,) Exit times (in lionrs) from tbe G2/M phase of tbe cell cycle in cnltnres of normal ( I ) and SV40-transformed melanocytes (A) treated with Apliidicolii> (cells blocked in G2/M). FACS analysis was performed in tbree independent sets of samples nsing cells of different donors. These are the results of o n e of tliree representative experiments.

SK-MEL 28 melanoma cells proliferated under these conditions with 30% plating efficiency. SV4() large T-antigen—transformed melanocytes were also examined for tuniorigenicity when injected into nude mice. In paixels of mice given normal melanocytes or T-antigen—positive melanocytes, there was no growth over a 5-nionth observation period, hi contrast, all five animals given SK-MEL 28 melanoma cells developed tumors at the site of inoculation within 14 d.

Fignre 5. T-antigen—expressing melanocytes acquire the ahility to synthesize hFGF. Each lane of a 12% PAGE-gel was loaded witb 30 jUg soluble intracellular protein obtained by detergent lysis (1% Triton X-100) of 5 X 10'' cells. After electrotransfer to a PVDF membrane (Inimobilon P), tlie proteins were stained witb bFGF antibody and visualized using tbe DAKO AEC-.substrate system. Lane I, SK-MEL 28 cells. Lane 2, SV40 T-antigen—transfoniied melanocytes. Lane .?, normal adult melanocytes.

cell cycle. This was accompanied by accelerated doubling times and a significantly extended life span iit vitro. In addition, in striking contrast to tlieir normal counterparts, these cells showed reduced dependence on bFGF. Expression of the SV40 large T antigen alone, however, proved not to be sufficient to fully establish the

% increase

DISCUSSION In the present study, the SV40 large T-antigen sequence was used to evaltjate the growth characteristics and malignant potential of human melanocytes isolated from normal skin. The rationale for tliis line of investigation was based on observations by other laboratories tbat melanoma susceptibility could be conferred by introduction of the SV40 transgene into melanocytes of transgenic mice [19,20]. However, even with this manipulation, conversion of melanocytes to melanomas was an infrequent event unless additional environmental factors such as ultraviolet radiation [19] or wounding of the skin [25,26] were added. The inability of SV4() to completely transform melanocytes to their malignant counterpart raises the question as to what it does do in a functional sei:isc to facilitate this process. We reasoned that valuable infonnation could be obtained by examining its effect on human melanocyte growth and progression through the cell cycle. Genomic integration of SV4() large T antigen caused significant alterations in the growth hehavior of these specialized cells. Shifts in proliferation kinetics iticreased the percentage of mitotic cells and decreased the duration of melanocytes in the GI phase o f the

* normal human melanocytes * T-ag positive melanocytes * melanoma line SK-MEL 28

0.2

0.4

O.I

0.8

ng/ml bFGF

Figure 6. T-antigen-transformed melanocytes require significantly less hFGF to proliferate. Effect of various bFGF concentrations on tbe proliferation of normal, SV40 T-antigen—transfoniied melanocytes and the SK-MEL 28 melanoma line. Cells were plated at a density of 4 X 10* cells/cm" in complete medium containing 0—1 ng/ml bFGF. After 10 d, cells were trypsinized and counted on a liematocytonieter. Tlie increase or decrease in cell numbers over tbe initial seeding concentration is depicted. • , T-antigen transformed melanocytes. • , Normal melanocytes. H. SKMEL 28 nielanom;! cells.

760

ZEPTER ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 7. Effect of various hFGF concentrations on the growth of normal a n d SV40 T-antigcn-transformed melanocytes and the SK-MEL 28 melanoma line. Cells were plated at a density of 4 X lo' cells/cm^ in complete m^^dium containing (a, d) no bFGF. b, c) 0.4 ng/ml bFGF. e,J) 1 ng/ml bFGF. Pictures were taken after 10 d of culture under selective conditions. Top, normal melanocytes. Bottom, T-antigen expressing. Pbase-contrast pbotomicrograpbs, bars, 50 jxm for a and d; 15 /J-m for b, c, e and /

malignant phenotype. Transformed melanocytes failed to form colonies in soft agar, they were unahle to initiate tumors in nude mice, and they eventually underwent senescence. These findings in in vitro transformed human melanocytes are compatible with those obtained in vivo in transgenic mice in wliich the SV40 transgene was specifically targeted to melanocytes [19,20]. In those animals, an increase in the number of proliferating

mclatjocytes was detected in vivo |19], and, when they were retHoVed and placed in culture, they exhihited an increased proliferative activity [20]. Apart from putative growth-stimulatory factors associated with wound healing, the administration of a single doSe of 0.7 mj/cni^ UVB—non-tumorigenic for normal cells— conferred anchorage-independence and, with slightly higher doses, a malignant phenotype to melanocytes from these transgenic mice.

VOL. 104. NO. 5

SV40 LARGE T-ANTIGEN-TRANSFORMED MELANOCYTES

MAY 1995

The SV40 large T antigen has added a new dimension to the oncogene-centered model of human carcinogenesis, in large part becatise of its ability to inactivate the protein product of the p53 tumor suppressor gene 115,27]. Stable complexes are fonned between tbe SV40 large T antigen and with highly conserved regions of the suppressor protein that exert specific growth-regulatory funcdons [15,27,28]. These domains contain the mutational "hot spots" found to be the most common genetic alterations in neoplastic development [29]. To date, binding of T andgen to p53 is considered to be a crucial step in the initiation of the transfoniied phenotype in non-peniiissive cells [29—32]. It thus seems reasonable to hyjiothesize that SV40 large T antigen mediates its effects on melanocytes largely throtigh its interactions with p53. Several immunohistochemical [33-36] as well as molecular biologic approaches [35,37] provide evidence that p53 loss of function is involved in the development of malignant melanoma. However, significant controversy exists regarding the timing of p53 overexpression during this multistep process. Whereas sttidies by Ak.slen and Morkve [33] and Stretch et al [35] reveal p53 expression in a majority of primary melanomas (70% and 97%, respectively), thus indicating an early event in the tumor evolution, other authors [34,37] observe increasing p53 immunoreactivity or polymerase cbain reaction—detectable mutations witb more advanced stages of the disease. Our data would suggest that alterations in tumorsuppressor genes play a critical role early in melanoma development. However, it is clear that a dynamic biologic process as complex as tbe neoplastic microevohitioii needs additional functional studies to evahiate the relevance of existhig experimental data. In normal cells, elevated levels of p53 during GI indicate a "checkpoint" phase of DNA control allowing the repair of potential mutations or the induction of apoptosis in cases of irreparable damage [14,38]. In our system, much faster progression of SV4() large T-antigen-transformed melanocytes through this stage of the cell cycle suggests the lack of active p53, a consistent finding in other T-antigen—expressing cultures as well [21]. In addition, transformation with the SV40 large T antigen stimulated quiescent cells to re-enter the cell cycle, an effect reflected by a higher percentage of mitotic cells (G2 peak) and smaller GO/GI fraction in our flow-cytometric analysis. SV40 large T-antigeu complexes with at least four other cellular proteins, including DNA polymerase a'*'', the product of the Rb-genell, a 73-kd heat shock—like protein [40] and another cellular protein ranging from 107 to 120 kd [41]. However, various experiments with truncated [31] or mutated [9,12] T proteins have clearly deinonstrated that an intact p53-binding .site is indispensable for tbe transforming effect of the simian virus oncogene [32]. Autocrine production of bFGF is a sensitive marker for nielanocytic cells acqtiiring malignant potential [23,24,42]. Melanoma cell lines proliferate withotit the addition of this melanocyte mitogen to cell cultures, and in melanocytes an increasing ability to synthesize bFGF parallels the capacity of those cells to grow in a semisolid milieu [18]. Further support for the relationship between bFGF and the malignant potential of melanocytes has come from studies in which an anti-sense bFGF construct was introduced into hunian melanoma cells. Such an experimental manipulation inhibited tbe proliferative activity of those cells and retarded their anchorageindependent growth in semi-solid medium [42]. Consistent with the hypotheses generated from those studies, our SV40 large T-antigen-positive melanocytes not only maintained their growth dependence on hFGF but also remained anchorage dependent. This would suggest that SV40 large T-antigen—transformed melanocytes are at an intermediate stage between nomial and malignant cells. Large T-antigen-expressing melanocytes remained morphologically indistinguishable from non-transformed cells, indicating tbat the SV40 large T antigen confers predominantly growth-related alterations on melanocytes. This contrasts to the findings of Jambrosic, who reported significant changes in melanocyte morphology upon T-antigeu transformation [17] with loss of dendrites and epithelioid growth pattern. However, iti that study, cells were

761

selected hy proliferation in the absence of PMA, which is known to induce dendrite formation [43] and might explain differences in the two studies. Maintenance of the original morphology has also been observed in p53-negative osteosarcoma following transfection of a mutant p53 gene [44], thus indicating that its loss of function is not related to cellular appearance. We also found that contact inhibition, another growth-regulatory factor, was also suppressed in large T-antigen-positive melanocytes. Growth to highly confluent monolayers has been observed in other similarly transformed lines [8—10,45]. This, too, has been attributed to interactions witb tbe p53 protein product. Defective mutants of T antigen lacking tbe binding site for p53 fail to enhance the sattiration density in transformed fibroblasts, whereas Rb-site deficient yet p53-intact T mutants induced saturation densities identical to wild-type T antigen [45]. In addition, transfection of Saos-2 cells with a mtitant p53 gene resulted in saturation densities that were up to five times greater than nonnal levels [44]. Transfer of wild-type forms significantly blocked tbe further growth of tumor cells [44]. In summary, the accumulated evidence suggests that by reducing the proportion of melanocytes in the GI phase of the cell cycle and by decreasing their dependence on exogenous growth factors, viral oncogenes such as the SV40 large T antigen predispose this specialized cell population to additional insults that eventuate in their conversion into fully malignant cells.

We wish to tliaiik Ms. Karen Tiibesiug and Ms. Hai-piiig Tang for tlieir expert technical assistance. Dr. James Jacobberger for helpful discussions, and Dr. David Kaplan for bis critical review if the manuscript. A. C. H. is tbe recipient of a research felloivsliip grant from the American Dermatology Foundation. K. Z. is the recipient of a felloti'ship grant Jrom tbe DFG, Gennany. Tbis work was supported by NIH giants AR397.W, AR.U593, CA4S735, CM576-JJ, and CA43703, and by a grant Jrom the Joints Hopkins Center for Alternatives to Animal Testing.

REFERENCES 1. 2. 3.

4. 5.

6. 7.

8. 9.

Albino AP. Fountain JW: Moleculnr genetics of human malignant melanoma. Cancer Treat Res 65:2(11-255. 19').i Lee WH: Tumor suppressor genes—the liopo (editorial). RiSEB ] 7(1(1):819. 1993 Bartek J, VojtcSL-k H. Grand RJ. Gallimorc I'H. Laiu- DP: Cellular localization and T antigen binding ot" the retinohlastoina protein. Oucoi^euc 7:1167-1175. 1992 Bishop IM: Molocular tbenu's in oncogenesis. Cell 64:23.5-248, 1991 Hnsain Z, FitzGciald GB. Wick MM: Comparison of cellular protooiicogenc activation and tratistbrniation. related activitj- of human meianocytes and uictastatic melanoma. / invest Deintatol 95:571-575. 1990 Privcs C. Bargonctli j . Friedman PN. Manfredi JJ. Wang EH: Functional consequences of the interactions ot the p53 tumor snppressor protein and SV40 large tnmor antigen. Cold Spriui; Harh Symp Quaiit Biol 56:227-235, 1991 Chen J. Tobiu GJ. Pipns JM. Van Dyke T: T-antigen mutant activities in vivo: roles of p5.3 and pRB binding in tumorigenesis otthe choroid plexus. Oiicogeiie 7:1167-1175. 1992 Goolsby CL. Steiner M. Nemeth J: Viral and cellnlar oncogene expression during progressive malignant transfomiation of SV40 transfonned human fibroblasts. Cylomelry 12:748-756. 1991 Kierstead TD. 1 evethia MJ: Association ot p53 binding and immortalization of primary C57BL/6 mouse embryo fibroblasts by using simian virus 40 T antigen mutants bearing internal overlapping deletion nintations. / I'iiol 67:1817-1829. 1993

10.

tl. 12.

13.

14.

Thompson DL. Kalderon D. Smith AE. Tevethia MJ: Dissociation of Rb. binding and anchorage, independent growth fi-oni immortalization and tnmorigenicit)nsing SV40 mutants producing N-terminally trnncated large T antigens. Firi%Y 178:15-34. 1990 Lane DP: p53 guardian of the genome. Niirwir 358:1.5-16. 1992 Dobbelstein M. Arthnr AK. Delide S. vau Zee K. Dickmanns A. Fanning E: Intracistronic complemeutatiou reveals a new timcdon of SV40 T antigen that co-operates with Rb and p5.^ binding to stimulate DNA synthesis in qniescent cells. Onf(>i;<-Mr 7:837-847, 1992 Li B, Fields S: Identification of mutations in p53 that atFect its binding to SV40 large T antigen by using the yeast two-hybrid system. FASEB J 7:957-963, 1993 Ln X. Lane DP: DitFerential induction of transcriptionally active p53 following UV or ionizing radiation: defects in chromosome instability syndromes? Cell 75:765-778, 1993

762

15. 16.

17.

18.

19.

20.

21. 22. 23. 24.

25. 26. 27. 28.

29. 30. 31.

THE JOURNAL OF INVESTIGATIVE DEIUVIATOLOGY

ZEPTER ET AL

Fanning E, Knippers R. Structure and function of simian virus 40 large tumor antigen. Atmu Rev Biochem 61:55-85, 1992 Jambrosic J, Mancianti ML, Ricciardi R, Sela BA, Koprowski H, Herlyn M: Transformation of normal human melanocytes and non-malignant nevus cells by adenovirus 12/SV40 hybrid virus. IntJ CoHCcr 44:1117-1123, 1989 Melber K, Zhu G, Diamond L: SV40-transfected human melanocyte sensitivity to growth inhibition by the phorbol ester 12-O-tetradecanoylphorbol, 13acetate. Cancer Res 49:3650-3655, 1989 Heriyn M, Clark W H , Rodeck U, Mancianti ML, Jambrosic J, Koprowski H: Biology of tumor progression in human melanocytes. Lab Invest 56:461-474, 1987 Larue L, Dougberty N, Mintz B: Genetic predisposition of transgenic mouse melanocytes to melanoma results in malignant melanoma after exposure to a low ultraviolet B intensity nontumorigenic for normal melanocytes. Proc Natl Acad Sci USA 89:9534-9538. 1992 Larue L, Dougherty N, Bradl M, Mintz B: Melanocyte culture lines from Tyr-SV40E transgenic mice: models for the molecular genetic evolution of malignant melanoma. Ottcogctte 8:523-531, 1993 Sladek TL, Jacobberger JW: Rapid titration of retroviral vectors encoding intracellular antigens by flow cytometry. J Virol 64:3135-3138, 1990 Kwon BS, Haq A, Kim GS, Pomerantz SH, Halaban R: Cloning and characterization of a human tyrosinase cDNA. Prog Clin Biol Res 256:273-282, 1988 Halaban R, Ghosh S, Baird A: bFGF is the putative natural growth factor for human melanocytes. In Vitro Cell Dev Biol 23:47-52, 1987 Herlyn M, Mancianti ML, Jambrosic J, Bolen JB, Koprowski H: Regulatory factors that determine growth and phenotype of nonnal human melanocytes. Exp CeU Res 179:322-331, 1988 Bradl M, Klein-Szanto A, Porter S, Mintz B: Malignant melanoma in transgenic mice. Proc Natl Acad Sci USA 88:164-168, 1991 Mintz B, Silvers WK: Transgenic mouse model of mahgnant skin melanoma. Proc Natl Acad Sci USA 90:8817-8821, 1993 Finlay CA: Normal and malignant growth control by p53. Cancer Treat Res 63:327-344, 1992 Lin JY, Simmons DT: The ability of large T antigen to complex with p53 is necessary for the increased life span and partial transformation of human cells by simian virus 40. J Virol 65:6447-6453, 1991 HoUstein M, Sidransky D, Vogelstein B, Harris CC: p53 mutations in human cancers. SaV/ia-253:49-53, 1991 Prives C: Doing the right thing: feedback control and p53. Cifrr Opin CeU Biol 5:214-218, 1993 Tevethia MJ, Pipas JM, Kierstead T, Cole C: Requirements for immortalization

32.

33. 34. 35. 36.

37.

38.

39. 40.

41.

42.

43.

44. 45.

of primary mouse embryo fibroblasts probed with mutants bearing deletions in the 3' end of the SV40 gene. Virology 162:76-89, 1988 Zhu JY, Abate M, IUce PW, Cole CN: The abiHty of simian virus 40 large T antigen to immortalize primary mouse embryo fibroblasts cosegregates with its ability to bind to p53. J Virol 65:6872-80, 1991 Akslen LA, Morkve O: Expression of p53 protein in cutaneous melanoma. IutJ CflHCcr 52:13-16, 1992 Lassam NJ, From L, Kahn HJ: Overexpression of p53 is a late event in the development of malignant melanoma. Caucer Res 53:2235-2238, 1993 Stretch JR, Gatter KC, Ralfkiaer E, Lane DP, Harris AL: Expression of mutant p53 in melanoma. Cancer Res 51:5976-5979, 1991 Vamamoto M, Takahaslii H: Immunohistochemical detection of the p53 oncoprotein in tumors of melanocytic origin. Virchows Arcliii' A Pathol Anat 422: 127-132, 1993 Weiss J, Schwechheinier K, Cavenee WK, Herlyn M, Arden KC: Mutation and expression of the p53 gene in malignant melanoma cell lines. IntJ Cancer 54:693-699, 1993 Kuerbitz SJ, Plunkett BS, Walsh WV, Kastan MB: Wild-type p53 is a cell cycle check point determinant following irradiation. Proc Natt Acad Sci USA 89: 7491-7495, 1992 Gannon JV, Lane DP: p53 and DNA polymerase alpha compete for binding to SV40 T antigen. Nature 329:456-458, 1987 I3utel JS, Jarvis DL, Maxwell SA: SV40 T antigen as a dual oncogene: structure and function of the plasma membrane associated population. Ann NY Acad Sci 567:104-112, 1989 Ewen ME, Ludlow JW, Marsilio E, DeCaprio JA, Millikan RC, Cheng SH, Paucha E, Livingston DM: An N-terminal transfomiation, governing sequence of SV4() large T antigen contributes to the binding of both pllORb and a second cellular protein, pl20. CeU 58:257-267, 1989 Becker D, Meyer CB, Heriyn M: Proliferation of human malignant melanomas is inhibited hy antisense oligodeoxynucleotides targeted against basic fibroblast growth factor. UMBOJ 8:3685-3691, 1989 pittelkow MR, Shipley GD: Seruni-free culture of normal hunian melanocytes: growth kinetics and growth factor requirements, y Cell Physiol 140:565—576, 1989 Chen PL, Chen YM, Book.stein R, Lee WH: Genetic mechanisms of tumor suppression hy the human p53 gene. Science 250:1576-1580, 1990 Su LN, Little JB: Transformation and radiosensitivity of human diploid skin fibroblasts transfected with SV40 T antigen mutants defective in Rb and p53 binding domains. IntJ Radiat Biol 62:461-468, 1992